Document 176569

eandhow t
Where and how to conserve:
Extending the scope of spatial reserve network design
Astrid J.A. van Teeffelen
Metapopulation Research Group
Department of Biological and Environmental Sciences
University of Helsinki
Academic dissertation
To be presented, with permission of the Faculty of Biosciences of the University of
Helsinki, for public criticism in the Auditorium of the Helsinki University Museum
Arppeanum (Snellmaninkatu 3) on the 16th of June 2007 at 12 o’clock noon.
Helsinki, 2007
© Astrid van Teeffelen (Summary, Chapter V)
© Authors (Chapters III, IV, VI)
© Springer (Chapter I)
© Oikos (Chapter II)
Cover design: Bart A. Smit & Astrid van Teeffelen
Layout: Astrid van Teeffelen
Author’s address:
Land Use Planning Group
Department of Environmental Sciences
Wageningen University & Research Centre
Droevendaalsesteeg 3
NL-6708 PB Wageningen
The Netherlands
E-mail: [email protected]
ISBN 978-952-10-3955-3 (paperback)
ISBN 978-952-10-3956-0 (PDF)
Printed by: Yliopistopaino
Helsinki 2007
Where and how to conserve:
Extending the scope of spatial reserve network design
Astrid J.A. van Teeffelen
The thesis is based on the following articles and manuscripts, which are referred to in the text by
their Roman numerals:
Van Teeffelen, A.J.A., M. Cabeza & A. Moilanen, 2006. Connectivity, probabilities and
persistence: comparing reserve selection strategies. Biodiversity and Conservation 15(3): 899–
Van Teeffelen, A.J.A. & O. Ovaskainen, 2007. Can the cause of aggregation be inferred
from species distributions? Oikos 116(1): 4-16.
Van Teeffelen, A.J.A. & A. Moilanen. Where and how to manage: optimal selection of
conservation actions for multiple species. Submitted.
Van Teeffelen, A.J.A., M. Cabeza, J. Pöyry, K. Raatikainen & M. Kuussaari. Maximizing
conservation benefit for grassland species with contrasting management requirements.
Van Teeffelen, A.J.A. Taking account of species-specific connectivity requirements when
selecting conservation actions. Manuscript.
Moilanen, A., A.J.A. van Teeffelen, Y. Ben-Haim & S. Ferrier. How much compensation is
enough? Explicit incorporation of uncertainty and time discounting when calculating
offset ratios for impacted habitat. Submitted.
Original idea
Study design
Methods &
Empirical data
*, **, ***
*, **, ***
AvT, MC, JP,
AvT Astrid van Teeffelen, AM Atte Moilanen, MC Mar Cabeza, OO Otso Ovaskainen, JP Juha
Pöyry, KR Katja Raatikainen, MK Mikko Kuussaari, JPy Juha Pykälä, SF Simon Ferrier, YBH Yakov
* Alterra - Wageningen University & Research Centre, the Netherlands.
** Province of Noord-Brabant, the Netherlands.
*** SWEV, West-Brabant Association of Bird Study Groups, the Netherlands.
Supervised by:
Dr. Atte Moilanen
Senior Researcher, University of Helsinki, Finland.
Reviewed by:
Dr. Risto Heikkinen
Senior Researcher, Finnish Environment Institute, Finland.
Dr. Kristina D. Rothley
Assistant Professor, Simon Fraser University, Canada.
Examined by:
Dr. Michael McCarthy
Senior Researcher, University of Melbourne, Australia.
Table of contents
1 Background
1.1 Where to conserve
1.2 How to conserve
2 Thesis outline
3 Results & Discussion
Spatial reserve network design
From a single to multiple conservation options in design
Connectivity in multi-option planning
Effects of uncertainy on fair offset ratios
4 Synthesis & Perspectives
Connectivity, probabilities and persistence: comparing reserve selection
Can the cause of aggregation be inferred from species distributions?
Where and how to manage: Optimal selection of conservation actions for
multiple species
Maximizing conservation benefit for grassland species with contrasting
management requirements
Taking account of species-specific connectivity requirements when
selecting conservation actions
How much compensation is enough? Explicit incorporation of uncertainty
and time discounting when calculating offset ratios for impacted habitat
About the author
Astrid J.A. van Teeffelen
Worldwide the impacts of human habitation
and economic development have resulted in
modification, destruction and fragmentation of
our natural environment (Hanski, 2005). As a
consequence, biodiversity is disappearing at an
alarming rate, which is 100 to 1000 times larger
than natural background rates of extinction
(Smith et al., 1993; Pimm et al., 1995; Pimm &
Assessment, 2005). This rate is expected to
increase even further due to human-induced
climate change (McLaughlin et al., 2002;
Thomas et al., 2004). In order to slow down the
rate of biodiversity loss, conservation efforts are
required. Habitat loss is considered to be one of
the main driving forces of species extinctions
(Saunders et al., 1991; Hanski, 2005). Preventing
further habitat loss is therefore a key issue if the
loss of biodiversity is to be slowed down. Due
to a high demand for natural resources, as well
as strong competition with other land use types
(like agriculture, urban areas and industry), the
pressure on natural areas is high. In
combination with the tight budget for
conservation, careful planning is required if we
want to preserve as much of biodiversity as
possible for future generations.
Reserves have been a means of protection of
natural heritage for centuries (Margules &
Pressey, 2000). The reasons for protecting areas
were mostly for their scenic beauty or for
recreational purposes such as hunting and
hiking, while biodiversity conservation was not
much of an issue (Pressey, 1994; Lindenmayer
& Burgman, 2005). Economic considerations
resulted in the location of most reserves at
economically uninteresting locations, such as at
high altitudes, steep slopes or on poor soils
(Adam, 1992; Pressey, 1994). These ad hoc and
biased reserve selection practises have lead to a
set of reserve networks that do not represent,
let alone protect, the full range of biodiversity
(Margules & Pressey, 2000). The insufficiency of
current reserves, the increasing pressure on
remaining natural areas and tight conservation
budgets motivated conservationists to develop
more systematic reserve selection methods for
designing reserve networks in a cost-effective
manner. The field of systematic conservation
planning that emerged as a consequence, covers
a wide range of questions about what facets of
biodiversity to focus our conservations efforts
on, where to conserve these biodiversity facets
and how to conserve them (Fig. 1). Here, I
mainly concentrate on the questions where and
how to conserve, see Box 1 for a short note on
what to conserve.
1.1 Where to conserve
In order to handle the question “where to
conserve?”, information is required on the
distribution of (surrogates of) biodiversity.
Henceforth, I will refer to “species” where I
mean to say “surrogates of biodiversity”. This,
because species are illustrative examples of
biodiversity, they are commonly used as
surrogates of biodiversity, and have been the
subject of study in this thesis. Other studies
have assessed the suitability of different species
groups as surrogates (see Box 1 for references).
Species distribution data
Species distribution data (e.g. from field
surveys or museum collections) are generally
classified as being either presence-only or
presence-absence information, and can be
binary or include abundance data (i.e. the
number of individuals at a location). Such data
are often incomplete and biased towards more
accessible areas (Polasky et al., 2000). Obtaining
a representative set of species distribution data
to inform conservation planning is timeconsuming, during which biodiversity loss
continues and options for conservation are
reduced (Wilson et al., 2005b).
The use of statistical models to predict the
occurrence of species at un-surveyed sites, is a
way to deal with incomplete distribution data
(Margules & Stein, 1989; Wilson et al., 2005b).
These models, referred to as species habitat
distribution models (SHDMs; or habitat
models) link species occurrence data to the
distribution of environmental variables, such as
vegetation, topographic and climatic variables
(Guisan & Zimmermann, 2000; Guisan &
Thuiller, 2005; Elith et al., 2006). Due to the
development of remote sensing technology, it is
relatively easy and cheap to obtain coverage
data for many environmental variables over
large areas, compared to obtaining complete
survey data for a large set of species. SHDMs
allow the fitting of these environmental
variables to the occurrence of a species, after
which the model can be used to predict the
probability of occurrence of that species at other
sites in the landscape, based on the
environmental suitability of these sites.
Methods for modelling species’ habitat
encompass – among others – generalised linear
models (GLMs), generalised additive models
(GAMs), and boosted regression trees (BRT)
(Guisan & Zimmermann, 2000; Elith et al.,
What to conserve?
How to conserve?
• Surrogates for biodiversity
• Goals
o Prioritizing species
o How much to conserve?
• Protection
• Restoration
• Maintenance
Where to conserve?
• Distribution of biodiversity
• Site selection algorithms
• Reserve network design
Figure 1. The three general topics (and a number of example subtopics) covered in systematic conservation
planning. The Roman numerals indicate how the chapters of this thesis are related to those.
Site selection problem:
Definitions & solution methods
In general, not all the locations where species
are found (or are expected to occur) can be
protected. Conservation planning therefore
requires decisions on which sites to select for
conservation. Constraints in conservation
budget or the number of sites (or total area) that
can be reserved have motivated the
development of site-selection methods, which
aim at finding efficient sets of sites to represent
species. Initially, site-selection methods were
scoring approaches that ranked sites according
to their species richness, and selected as many
sites from the top of the list until all species
under consideration were represented in the set
of selected sites (Wilson et al., 2005b). However,
to design entire reserve networks, the
properties of individual sites should be
assessed in relation to each other, rather than
individually – a concept known as
complementarity (Vane-Wright et al., 1991;
Margules & Pressey, 2000). Cost considerations
are an explicit element of site-selection
methods, although true cost of representation is
oftentimes approximated by the total number
or area of sites.
Two problem definitions are commonly used in
site-selection problems: 1) the minimum set
covering problem, which aims at representing
Box 1. What to conserve?
Although the goal of conservation planning is the protection of biodiversity, it is impossible to
obtain information on the distribution of all aspects of biodiversity (i.e. genes, individuals,
demes, populations, metapopulations, species, communities, ecosystems and their interactions;
Lindenmayer & Burgman, 2005) to direct conservation efforts. Instead, we need to rely on
biodiversity surrogates to represent the full extent of biodiversity (Pearson, 1994; Faith &
Walker, 1996; Margules et al., 2002). Potential surrogates encompass species assemblages, taxa
subsets, vegetation communities and environmental diversity (Ferrier et al., 2004; Chiarucci et
al., 2005; Trakhtenbrot & Kadmon, 2005). A large amount of literature is devoted to the testing of
surrogate groups for their ability to represent other aspects of diversity (Howard et al., 1998;
Andelman & Fagan, 2000; Williams et al., 2006). Although ideal surrogates probably do not
exist, quantification of what we aim to conserve is essential for systematic conservation
The debate on “how much conservation is enough”, is decades old, and still ongoing (e.g.
Rodrigues & Gaston, 2001; Svancara et al., 2005; Tear et al., 2005). For individual species of
conservation concern, modelling approaches such as population viability analysis (PVA) can be
used to estimate extinction risk and evaluate conservation scenarios (Burgman et al., 1993;
Morris & Doak, 2003). Such models are typically parameter rich and therefore data-demanding,
which complicates the application of PVA to large sets of species and large scale conservation
planning. At these large scales, more general rules of thumb are needed to determine practical
conservation goals. For example, The World Conservation Union advocates to set aside at least
10% of the land area of each nation for biodiversity conservation (IUCN, 1993), which has in
turn been criticized for being insufficient for long term conservation (Soulé & Sanjayan, 1998;
Svancara et al., 2005). Although there is consensus on the goals of conservation (the long term
protection of species and biodiversity in general), transforming broad goals into quantitative
objectives is essential for a decision-theoretic approach to the conservation problem (Westphal &
Possingham, 2003; Nicholson & Possingham, 2006).
each species a given number of times while
minimizing cost (Underhill, 1994; Csuti et al.,
1997; Pressey et al., 1997). The proportional
coverage problem is a variant of this, which
aims at representing a particular proportion of
vegetation types (Pressey et al., 1997; Leslie et
al., 2003; Moilanen & Cabeza, 2005); 2) The
maximal coverage location (or maximum
coverage) approach, which maximizes the total
number of species representations under a cost
constraint (Camm et al., 1996; Church, 1996;
Arthur et al., 1997; Arponen et al., 2005).
Instead of including simply the number of
representations, more meaningful measures of
population viability are being sought by e.g.,
using probabilities of occurrence rather than
presence-absence data (e.g. Araújo & Williams,
2000; Williams & Araújo, 2000, 2002; Wilson et
al., 2005b), or by including measures of viability
and persistence in the objective function
directly (Hof & Raphael, 1993; Bevers et al.,
1995; Araújo & Williams, 2000; Nicholson &
Possingham, 2006).
From site selection to reserve network design
Initially, site selection methods did not consider
the size or spatial configuration of the set of
selected sites. Consequently, a small and
spatially scattered set of sites was typically
selected, because such a set is inexpensive and
thus efficient. This would not be problematic, if
the sites remain embedded in a landscape
consisting of natural vegetation. However, with
ongoing land use change, reserves are likely to
become islands of natural habitat in a matrix of
intensive land use. In landscapes with small
and scattered habitat patches, local populations
are sensitive to demographic stochasticity due
to small population size and therefore have an
increased risk of extinction. Edge effects can
cause habitat quality at the edge of a patch to be
lower than at the core of the patch. Examples
of edge effects are the influx of high nutrient
levels or pesticides from neighbouring
agricultural land; lower availability of resources
such as food and shelter; a higher predator
density in the matrix, which increases
predation risk at the edge (Andren &
Angelstam, 1988; Woodroffe & Ginsberg, 1998;
Debinski & Holt, 2000). Since small patches
have relatively more edge habitat than large
patches, edge effects are more pronounced in
smaller patches.
Though local populations at small patches are
more likely to go extinct, a species could persist
in a fragmented landscape as a metapopulation,
if the exchange of individuals between patches
through dispersal is sufficient (Hanski, 1998,
1999). Dispersing individuals can found a new
population in empty patches, and reduce
extinction probability of existing populations
via the so-called rescue effect (Brown & KodricBrown,
individuals between patches is typically
reduced with increasing distance between
patches. Simply because the likelihood of
finding another patch is reduced, and also
because mortality risk is usually higher in the
matrix: due to higher predator density, the
presence of barriers such as roads and build-up
areas and a lower resource density.
Consequently, a lower dispersal success
decreases colonisation success and increases the
extinction probability of populations, which
metapopulation persistence. Species living in
larger, better connected networks therefore
have a larger probability of persistence than
species living in small, fragmented networks
(Hanski, 1998).
From these concepts of metapopulation ecology
follows that the size and the configuration of
networks should be larger and more compact in
order to increase the probability of species
persistence in the reserve network. Also from a
management perspective clustered reserve sites
can be preferential, as management costs
typically increases with increasing reserve edge
length (Possingham et al., 2000). The
approaches that take the spatial configuration
of reserve networks into account can be divided
into 1) methods that implicitly take account of
species’ connectivity requirements by aiming at
a structurally connected reserve network; and
2) methods that explicitly adapt the
configuration of the network to species-specific
requirements, hence selecting
connected reserve networks.
Structurally connected reserve networks can be
obtained by applying a penalty for the total
boundary length of the reserve network, which
causes a preference for selecting sites with a
clumped spatial configuration (Possingham et
al., 2000; Nalle et al., 2002; Cabeza et al., 2004b).
Other approaches to decrease the distance
between sites encompass methods that
minimize the maximum distance between sites
or the sum of inter-site distances (Briers, 2002;
Önal & Briers, 2002), and the selection of
adjacent sites for habitat restoration, which
together would form a corridor between
patches (Williams & Snyder, 2005). The interior
point search method introduced by Moilanen
(2005a) is yet another method, which selects a
given number of spatially contiguous reserves.
One potential disadvantage of forcing
structural connectivity into the network
configuration is that sites will be included that
have low habitat quality or few species, for the
sake of clustering alone. The reservation of such
sites requires resources, which precludes the
selection of other sites, which are perhaps more
valuable for conservation. Another drawback of
Box 2. Solution methods for site selection
Small site selection problems with few species can be solved by hand, but conservation
planning problems typically encompass large areas and many species, for which finding
efficient and effective solutions is no longer straightforward. Computer algorithms have
therefore been developed to aid the site selection for large problems, which can be divided into
exact and non-exact (or heuristic) methods. Exact methods, such as linear and integer
programming are guaranteed to find globally optimal solutions (Underhill, 1994; Church, 1996;
Rodrigues & Gaston, 2002; Williams et al., 2004), but have two limitations: 1) The objective
function and constraints need to be linear. When the value of a site depends on the value of
other sites in the set of selected sites, however, non-linearities arise (see last paragraph of
section 1.1); 2) Linear programming may run into computational difficulties for large reserve
selection problems (> ~ 10,000 selection units) due to the enormous increase in the search space
size (Williams et al., 2004). As conservation problems typically encompass many sites and
species, and are increasingly having non-linear problem formulations the use of exact
algorithms is not always feasible.
Non-exact methods encompass iterative heuristics (i.e. local search methods based on stepwise
maximization of marginal gain) and stochastic global search techniques. Iterative heuristic
algorithms are quick , easy to implement and hence widely used (Kirkpatrick, 1983; Pressey et
al., 1993; Pressey et al., 1996; Williams et al., 1996; Araújo & Williams, 2000), although they have
been criticized for returning suboptimal solutions (Underhill, 1994; Rodrigues & Gaston, 2002;
Önal, 2003). Stochastic global search techniques, such as simulated annealing (Possingham et
al., 2000; McDonnell et al., 2002; Westphal & Possingham, 2003; Westphal et al., 2007) and
genetic algorithms (Moilanen & Cabeza, 2002; Moilanen, 2005a, b) can be demonstrated to find
near-optimal solutions for large, complex (e.g. non-linear) optimization problems. Whether or
not solutions from site selection algorithms are mathematically optimal is no longer a central
debate in conservation planning, as conservation is more than the output of an algorithm
(Pressey et al., 1996). Moreover, the effects from incomplete data, model assumptions and
problem formulations are likely to have a larger influence than an optimal versus a slightly
suboptimal site selection solution.
these methods that work directly on the spatial
pattern of reserve sites is that they do not
account for species-specific requirements for
spatial configuration. As species differ in their
habitat requirements and dispersal capacity, the
optimal configuration of reserve networks will
be different for different species.
To explicitly account for species-specific
parameterize a species habitat distribution
model that incorporates a measure of
connectivity as one of the covariates, resulting
in a spatial SHDM. The neighbourhood over
which the connectivity is calculated should be
related to the species requirements for
connectivity, which could come from e.g.
dispersal capacity, sensitivity to edge effects or
home range size (see also Box 3). Examples of
spatial SHDMs are found in Ferrier et al. (2002),
Binzenhofer et al. (2005) and Wintle et al.
(2005). Such spatial SHDMs can next be
incorporated in reserve selection, to account for
species’ connectivity requirements through the
predicted species habitat distribution (Cabeza,
2003; Westphal & Possingham, 2003; Cabeza et
al., 2004a; Moilanen, 2005b; Moilanen et al.,
2005; Moilanen & Wintle, 2006, 2007; Westphal
et al., 2007). As a result of this, the model will
predict better connected sites to have higher
occupancy levels than equally suitable, but less
well connected sites. The reserve selection
algorithm will prefer sites with high predicted
occupancy levels, and will thus select better
connected sites.
Besides accounting for the spatial configuration
in reserve sites, it is also important to account
for the possibility of landscape change around
reserve networks. As metapopulation theory
predicts, the occurrence of nearby populations
affects the probability of occupancy of a site
(Hanski, 1998). When selecting reserve
networks, it is therefore important to
acknowledge the possibility of habitat loss
outside reserve sites, to prevent the selection of
small sink sites, where the occurrence of species
depends on source populations outside the
network (Cabeza & Moilanen, 2003). One way
to do this is by letting the predicted
probabilities of occurrence within reserve sites
depend explicitly on the set of reserve sites
only, as was introduced by Cabeza (2003) under
the concept of the dynamic probability approach
(see also Moilanen, 2005b; Westphal et al., 2007;
Chapters I & V). This approach requires the
probabilities of occurrence to be re-computed
during the site selection process, to account for
changes in the set of selected sites. Note that
this results in a non-linear problem formulation
(Box 2), which cannot be solved by exact
solution methods. Other approaches that
explicitly account for habitat loss around
reserve networks are described by Moilanen et
al. (2005) and Moilanen & Wintle (2006; 2007).
1.2 How to conserve
Though the protection of remaining natural
habitat is vital for biodiversity conservation, it
is not the only conservation practise (Fig. 1;
How to conserve). Especially in humandominated landscapes around the world vast
areas of natural habitat have been lost or are
degraded (Hanski, 2005; Moilanen & Wintle,
2007). Protecting only remaining habitat may
not be sufficient for conservation if there is too
little high quality habitat available for species to
persist in the long term. Habitat restoration can
be defined as the management of degraded
sites, with the aim to improve site quality in
terms of biodiversity (Gilbert & Anderson,
1998). By restoring degraded sites, or applying
agri-environment schemes, the conservation
value of a landscape can be improved, which
makes habitat restoration a complementary tool
for biodiversity conservation (Dobson et al.,
1997; Young, 2000; Donald & Evans, 2006). Next
to restoration and protection measures, also
maintenance and management of reserved or
restored sites may be required to maintain
optimal site conditions for biodiversity
conservation (Margules & Pressey, 2000).
As described above, the field of reserve design
has a relatively long tradition in the use of
optimization tools to select optimal reserve
networks for a large number of species in a
cost-effective manner (Margules & Pressey,
2000; Cabeza & Moilanen, 2001; Williams et al.,
2004). The field of restoration planning has a
concentrating on a single species or habitat type
for which a number of land use planning
scenarios are evaluated. Population viability
analysis (PVA) is a frequently used method to
evaluate restoration scenarios for a single
species (Root, 1998; Haines et al., 2005;
Schtickzelle et al., 2005). Other single-species
restoration planning approaches aim at ranking
sites by restoration value (Nikolakaki, 2004;
Schultz & Crone, 2005). The application of
optimization algorithms to restoration planning
problems has however been limited (Hof et al.,
2002; Newbold & Eadie, 2004; Crossman &
Bryan, 2006; Westphal et al., 2007). This is
surprising, as restoration planning problems
can be formulated like reserve selection
problems (Westphal & Possingham, 2003),
which allows for (near-) optimal planning
solutions rather than the comparison of a few
scenarios only.
Towards multi-action conservation planning
Reserve selection problems can be considered
single-action conservation planning problems,
as the application of a single conservation
action (i.e. protection) is considered per site.
Restoration planning problems that consider a
single action (e.g. revegetation), fall also into
this category. Many conservation planning
problems however, are characterized by the
availability of multiple alternative conservation
actions per site. For example, the restoration
and maintenance of grasslands or moorland can
require methods like grazing (with varying
intensity), sod cutting or prescribed burning.
Also reserve planning problems could have
multiple alternative conservation actions per
site, for example when planning different
protection or zoning levels (e.g. strict reserves,
buffer areas, and recreation areas). Despite the
apparent commonness of such conservation
planning problems, methods in the field of
reserve network design have not been adjusted
conservation actions. A few studies related to
forest harvest scheduling and landscape
planning have considered multiple alternative
actions per site (Hof et al., 1994; Bevers et al.,
1995; Holzkämper et al., 2006). However, these
studies did not consider the effects of action
costs and budget availability on species
representations. In contrast, cost constraints
motivated the development of systematic
conservation planning tools and are hence a
core element of systematic planning tools as
used for reserve selection.
Restoration as compensation measure
Although the protection of natural vegetation
needs to be preferred above restoration as a
conservation tool (Young, 2000; Hilderbrand et
al., 2005), the degradation or loss of existing
habitat cannot always be prevented, for
infrastructural works are considered necessary.
compensation for the loss through restoration
of habitat elsewhere is oftentimes required by
law (Ten Kate et al., 2004). Restoration activities
to compensate for the loss of biodiversity at the
impacted site are generally referred to as
biodiversity offsets (Ten Kate et al., 2004). Besides
the importance of planning where and how to
restore, the main question in this context is
“how much to restore?” (Fig. 1) (Race &
Fonseca, 1996; Bakker et al., 2000). Quantifying
the value of the lost and to-be-restored habitat
is hard, and hence defining offset ratios is
difficult. Whether restoration practises will
result in the desired outcome for biodiversity
remains to be seen, due to e.g. uncertainties in
restoration action success and establishment of
the species community, and time delays
(Hilderbrand et al., 2005; Morris et al., 2006;
Roach & Wade, 2006). Habitat Equivalency
Analysis (Dunford et al., 2004; Bruggeman et
al., 2005) and the concept of “No Net Loss”
(Harper & Quigley, 2005) are often used in this
context to aim at sufficient compensation.
However, none of the methods incorporates the
various sources of uncertainty in finding
appropriate offset ratios, for which the question
of “how much to restore?” remains largely
Species distribution data
Environmental variables
Understanding species
distribution pattern
Species habitat distribution model
I, II, IV, V
Predicting effects
of conservation
Site (and action)
selection algorithm
Prioritising species
Valuing sites
and actions
Goal setting
Selecting sites and
Implementation of conservation plan
Figure 2. A schematic overview of where the topics of this thesis (in dark grey boxes, referred to by their
Roman numerals) are placed within a conservation planning framework. The light grey box depicts an
algorithm for the selection of sites (chapter I) and actions (chapters III, IV &V).
This thesis relates to several aspects that come
into play in the planning of conservation effort
such as reservation and restoration (Fig. 2). The
chapters in this thesis can be linked in different
ways. Three chapters concern spatial issues in
the planning of conservation effort (I, II, V). I
will first introduce chapters I and II, chapter V
will be introduced together with chapters III
and IV in the context of multi-action
conservation planning, followed by chapter VI
which concerns the estimation of robust
biodiversity offset ratios.
Spatial issues in planning conservation actions
Since species differ in habitat requirements and
dispersal abilities, the optimal location and
configuration of reserve networks is expected to
be different for different species. As
conservation efforts cannot be targeted to each
species individually, solutions are sought that
are efficient in terms of area and cost, while still
aiming at protecting the species. To aim for
more robust reserve networks in terms of
species persistence, measures that are expected
to correlate with species viability have been
incorporated in site selection methods. For
example, using probabilities of occurrence
rather than presence-absence data, and
enhancing connectivity of reserve networks. In
chapter I is investigated how the size and
spatial configuration of reserve networks
changed with different input data (presenceabsence data and probabilities of occurrence
from both non-spatial and spatial SHDMs),
when assuming 1) a static landscape or 2)
complete habitat loss beyond the reserve
network. The different reserve networks were
evaluated in terms of the predicted occurrence
levels of species inside the reserve networks,
while accounting for habitat loss.
Species requirements for connectivity can be
estimated from the level of aggregation in
species distribution data. However, different
factors (e.g. dispersal, disturbances) can cause
similar levels of aggregation in species
distribution data, which may call for very
different spatial designs of ecosystem networks
(Box 3). Hence, it is important to investigate
what factors drive the spatial distribution of
species. Existing methods that account for
aggregation in species distribution data, such as
autoregressive models (Lichstein et al., 2002),
have not been used to examine what factors
caused aggregation in the species distribution
data. In chapter II is investigated whether one
can distinguish between different potential
causes of aggregation through the use of
different spatial variables in SHDMs. Such
information would be valuable to conservation
planning, as the spatial configuration of reserve
networks can be adjusted depending on the
variables that drive the spatial distribution of
From a single to multiple conservation options
Conservation actions such as habitat protection
or restoration are carried out in an attempt to
improve habitat conditions for species that
might not persist without taking action.
Planning of one such conservation action
therefore focuses on these species that are likely
to benefit from it. As action can not be taken at
all locations, the problem concentrates on
which locations are best targeted for
conservation, in order to maximize the gain for
the set of species of interest. Hence, the main
question dealt with in conservation planning
thus far is “where to conserve” (Fig. 1). When
considering alternative conservation actions per
site, the question extends to “how to conserve”
(Fig. 1). Planning of alternative actions at sites
can be relatively straightforward, when all
species of concern have similar requirements.
For example, planning which protection level to
apply to sites (e.g. recreational area, national
park, strict reserve), can be expected to have
similar effects on species (stricter protection
encompasses less disturbance, hence higher
occurrence probabilities for species vulnerable
to disturbance). The expected representation of
species under different actions (such as
protection levels) can be predicted with a
SHDM that incorporates the variables that will
change with different actions.
When species have contrasting requirements
however, a planning problem with multiple
actions becomes more difficult, as certain
actions can benefit particular species, but could
be harmful to other species. Effects of actions
can vary with the locations where they are
applied, as site conditions are usually not
homogenous over space. In order to make
decisions on which actions are best applied to
which sites, the effects of different actions on
species representation have to be valued.
Benefit functions can be used to translate the
changes in species representation, as a result of
applying an action, into a conservation value
(Hof & Raphael, 1993; Arponen et al., 2005;
chapter III).
In chapter III conservation planning problems
are defined with multiple actions per site, from
which the single-action selection problem (such
as reserve selection) can be considered a special
We describe a site-action selection
algorithm for multiple species and explore
ways to value effects of actions on species
representation, which could have a strong effect
on the outcome of selected conservation actions
(III, V). Chapter IV contains an application of
the method described in chapter III, to
investigate the trade-offs that can occur when
Box 3. Causes of aggregation in species distributions
Species commonly show an aggregated spatial distribution. The aggregation in species
distributions causes observations at sites that are close together to be spatially dependent on
one another, while most statistical methods require observations to be independent (Besag,
1972; Legendre, 1993; Guisan & Thuiller, 2005). To correct for this, autoregressive models have
been developed, which include a variable that measures the amount of occupied sites in the
neighbourhood of a site (Besag, 1972; Lichstein et al., 2002). This autocovariate resembles the
connectivity measures typically employed in metapopulation modelling (Ferrier et al., 2002;
Moilanen & Nieminen, 2002). The importance of connectivity for dispersal and metapopulation
persistence motivated the use of spatial covariates in species habitat distribution models
(spatial SHDMs) for reserve design (Cabeza et al., 2004a).
Dispersal limitation is an endogenous cause of aggregation, which relates to the behaviour of
the species, just like gregarious behaviour for example (Augustin et al., 1996). Exogenous
factors that can cause aggregation in species distributions include spatially aggregated
environmental variables such as climatic, soil or vegetation variables, regional stochasticity
(Hanski, 1991), and the occurrence of other species such as predators, prey or competitors.
These different factors can lead to similar levels of aggregation in species distributions.
Spatial covariates in SHDMs can capture spatial aggregation in species distributions but do not
distinguish between different processes that can cause aggregation. Some of these processes
however, call for a reserve network design that is dispersed (for factors like regional
stochasticity, disease, predators, fire) rather than aggregated, in order to sustain species at the
long term. To adjust reserve configurations to species’ spatial requirements it is important to be
able to disentangle what variables drive aggregation of species distributions. Thus far,
autoregressive models have not been used to investigate what factors caused the aggregation in
species distribution data.
seeking appropriate sets of sites and
conservation actions to benefit a set of species
with contrasting requirements. We examine the
robustness of the set of selected actions to
budget constraints and the relative priorities
assigned to species.
Where chapter I concerns the design of a
reserve network only, chapter V allows for
multiple alternative conservation actions per
site (Fig. 3). I describe a method to explicitly
account for a species’ habitat configuration
requirements, when selecting among multiple
protection and restoration measures per site.
Effects of uncertainty on fair offset ratios
Biodiversity offsets are a policy instrument to
compensate for habitat loss due to economic
development. To prevent net loss of
biodiversity value, an answer is required to the
question of how much restoration is enough to
compensate the loss of biodiversity value at the
impacted area. Several sources of uncertainty
are associated with the expected conservation
outcome of habitat restoration (Suding et al.,
2004; Hilderbrand et al., 2005). Chapter VI
addresses the influence of such uncertainties on
the calculation of fair offset ratios that are
unlikely to result in a net loss of biodiversity
value. Since restoration is increasingly used to
compensate habitat loss because of economic
development activities, it is increasingly
relevant that restoration practices result in
equal biodiversity value at the restored sites as
was lost from the developed sites. In this
chapter we investigate how much restoration
effort is required to match the realised value at
restored sites with the value lost from
developed sites, when taking into account that
restoration failure at particular sites is possible.
3.1 Spatial reserve network design
Habitat loss and connectivity
Selecting reserves based on SHDMs that
included spatial variables, resulted in more
aggregated reserves than when selection is
based on non-spatial SHDMs or presenceabsence data directly (Chapter I). These more
probabilities of occurrence inside reserve
networks, when habitat outside the reserve
network would be lost completely. Explicit
accounting for habitat loss at unprotected sites
during the site selection process, resulted in the
selection of larger and better connected reserve
networks. This occurred, because the
probabilities of occurrence of a species within
the reserve network decrease as a result of
habitat loss. Consequently, if we aim at
protecting species for which habitat is likely to
degrade or disappear without protection, we
have to set higher targets than for species for
which habitat is less likely to degrade or
disappear. The assumption of 100% habitat loss
landscapes this is not implausible as the
expansion pressure from urban areas and
industry is large. If preferred, land use change
modelling could be employed to provide other
scenarios of habitat loss.
Disentangling causes of aggregation
Several exogenous and endogenous processes
and variables can cause aggregation in species
distributions (Box 3). Chapter II investigated
the possibility to disentangle causes of
aggregation from snapshots of species
performance of spatial SHDMs with different
spatial covariate structures. Species distribution
data with similar aggregation patterns were
simulated with different causes of aggregation
(distance dependent dispersal, edge effects and
aggregated distributions of environmental
factors, which were static as well as dynamic in
space and time). Next, different spatial SHDMs
were fitted to these data, varying in the type of
spatial covariate. We observed differences in
the relative model fit of various spatial SHDMs,
which can be understood based on the cause of
aggregation that we used to model simulated
data. We also observed differences in model fit
of various spatial SHDMs when analysing
spatial occurrence patterns of several bird
species, indicating that information can be
gained from species distribution patterns in this
way. The use of spatial SHDMs in this manner
provides a possibility to study what variables
drive a species distribution, to aim for
biologically more meaningful habitat models.
More meaningful models will in turn be better
predictors of species occurrence, and hence be
more informative to conservation planning.
3.2 From a single to multiple
conservation options in design
In three chapters of this thesis I allowed for
multiple alternative conservation options per
site, instead of just one option as has hitherto
been employed in the field of reserve design
and in the few earlier applications of reserve
selection tools in restoration planning. In
chapters III, IV and V conservation actions
were valued based on the change in the
conservation value of the landscape (summed
over all species considered) relative to action
cost, in order to decide which action returns the
highest marginal gain in benefit. The shape of
the benefit function had a large impact on the
capacity of a forward stepwise heuristic
algorithm to find globally optimal solutions
(III). We showed that reserve selection
formulations with either concave or convex
benefit functions properties can be optimally
solved (III) even with a simple iterative
algorithm. As the computation time of the
iterative algorithm is short, this approach could
successfully be used for large conservation
problems with many sites, species and actions.
It is important to note however, that the choice
for a particular benefit function influences the
set of actions and sites that is selected by the
algorithm, and its consequences for the
expected representation of the species in the
landscape. Although all benefit functions aim at
maximizing the value obtained by conservation
actions, they value changes in representations
differently depending on the level of
representation (Hof & Raphael, 1993; Arponen
et al., 2005; III, IV). Arponen et al. (2005)
demonstrated that selecting reserve networks
with different benefit functions results in rather
different representation levels of for example
rare species. Where Arponen et al. (2005)
studied how benefit functions affect the
selection of reserve sites, this was extended
here to the selection of both sites and actions.
Even when optimizing conservation actions for
a single species, the shape of the benefit
function had a large effect on which actions are
preferred by an algorithm (V). Which benefit
function is considered better is not
straightforward to answer, as it depends on the
characteristics of the planning problem at hand
and the aim of the study (III, IV).
When the group of species in need of
conservation actions has diverse habitat
requirements, the species are likely to benefit
from different actions. For example, with the
disappearance of semi-natural pastures due to
changes in agricultural practices, preventing
succession through restoration is required to
benefit the species that depend on such
pastures. The species can differ however in how
intensive restoration is preferred (Köhler et al.,
2005; Pöyry et al., 2005; Pykälä, 2005; IV). The
planning of actions to cater for these different
species can be handled with the same problem
formulation and algorithm as presented in
chapter III. The algorithm selects an action at
each site, to maximize the summed benefit over
all species, while accounting for action cost. As
a result of species’ contrasting requirements
however, the level at which individual species
were represented differed widely when all
species were given equal priority. Although the
average representation over all species did not
appear very sensitive to the budget available,
the individual species representations differed
substantially, indicating a strong trade-off
between different species groups (IV). The
results from chapter IV emphasize the
importance of using sensitivity analysis to test
the robustness of conservation planning
solutions to the relative priorities assigned to
species, when planning for species with
different requirements. Consequently, on top of
the importance of the choice for a particular
benefit function to value changes in species
representation levels as a result of different
actions (III), it becomes also important to
determine how to prioritize between different
species (IV, Fig. 1).
3.3 Connectivity in multi-option
The method presented in chapter III and
applied in chapter IV is non-spatial. It is
possible though to employ for example a
penalty for boundary length or for distance to
connectivity into the set of selected sites and
actions, like in single-action planning problems
(e.g. Possingham et al., 2000; Cabeza et al.,
2004a; Crossman & Bryan, 2006; I). An
approach that incorporates functional (i.e.
species-specific) connectivity in planning with
multiple conservation options per site for a
single species, is presented in chapter V (Fig. 3).
The algorithm, which is a reverse stepwise
heuristic, starts from the ‘ideal’ landscape for a
species by assigning the best action at each
location, followed by the downgrading of
actions at sites, until the remaining set of
actions can be afforded. When selecting which
actions to replace by cheaper (but less
beneficial) actions, the effect of replacing an
action is calculated for the local site, as well as
for the neighboring sites, through a spatial
SHDM, similar to the dynamic probability
approach described by Cabeza (2003) for a
single-action problem formulation. Hence, the
effects of restoration and protection actions are
requirements for habitat and connectivity.
Again, the shape of the function used to
evaluate the effect of actions on the species
representation had a large effect on the priority
assigned to actions. The method as presented in
chapter V works well for a single species with
simple habitat requirements, and could also be
applied to multiple species that respond
similarly to particular actions (as in the
example of different protection levels with
increasing strictness). For species with
responses to actions can be contradictory, for
which the ‘ideal’ landscape to start the action
removal from, is not easily defined.
3.4 Effects of uncertainy on fair
offset ratios
Figure 3. Classification of conservation planning
problems, and how the chapters of this thesis (in
Roman numerals) relate to this.
One way to determine the offset ratio for
compensation of a damaged site is by dividing
the present conservation value of the site that
will be lost, by the expected conservation value
of the sites to be restored. This approach, which
is termed ‘matching of mean utility’ in chapter
VI, ignores uncertainty and time lags in the
establishment of conservation value at
restoration sites. If the realized conservation
value at restored sites turns out lower than
expected, this results in a net loss of
biodiversity value, whereas the objective of
offset legislation is the maintenance of
biodiversity (Ten Kate et al., 2004). Instead of
matching mean utility, chapter VI employed
uncertainty analysis, to calculate so-called
‘robustly fair offset ratios’ (Fig. 2). These offset
ratios are robust in the sense that they
guarantee a high enough probability that the
compensation does not result in net loss of
biodiversity value. We accounted for several
factors which are likely to affect the expected
value of restored sites: (i) uncertainty in the
effectiveness of restoration action; (ii)
uncertainty about the establishment and
growth of conservation value at compensation
areas; (iii) correlation between success of
different compensation areas; and (iv) time
discounting. When accounting for these
possible causes of restoration failure, the offset
ratio required to obtain at least the same
conservation value in the restored sites as
currently present in the development site,
increased steeply. The key point of this paper is
that higher offset ratios are needed for a
robustly fair exchange between destroyed and
restored habitat, when the aim is to have a low
probability that the exchange turns out
unfavourably. For maintaining biodiversity
value in the long run, the calculation of offset
ratios requires explicit accounting of various,
but common, sources of uncertainty.
In the struggle for biodiversity conservation in
a world where natural habitat becomes sparse
and conservation budgets are tight, methods for
systematic conservation planning have come a
long way (Margules & Pressey, 2000; Cabeza &
Moilanen, 2001; Williams et al., 2004). Besides
continuous efforts on delineating appropriate
conservation goals and biodiversity surrogates
(What to conserve, Fig. 1), reserve network
design methods are increasingly dealing with
direct and indirect measures of species viability
to increase the likelihood of persistence (Gaston
et al., 2002; Wilson et al., 2005a; Moilanen &
Wintle, 2006; Nicholson et al., 2006). With the
aim of long-term conservation on the one hand,
and data limitations on the other, trade-offs
have to be made with respect to model
complexity, and applicability of methods at
large scales for many species (Verboom et al.,
2001; Vos et al., 2001; Opdam et al., 2003;
Moilanen et al., 2005). Within this trade-off,
species habitat distribution models (SHDMs)
can be a means to overcome the problem of
incomplete survey data to inform conservation
planning, by predicting species occurrence
probabilities based on the habitat suitability of
sites, which has been considered a measure of
viability (Araújo & Williams, 2000; Williams &
Araújo, 2000; but see Wilson et al., 2005b).
SHDMs also allow for the inclusion of spatial
covariates (also termed connectivity measures
or contextual variables), which can be useful for
several reasons demonstrated in this thesis: In
the context of selecting a network for
conservation, spatial SHDMs allow for the
evaluation of configuration of reserve networks
and other conservation actions for either
structural or functional connectivity (I, V).
Spatial SHDMs also facilitate incorporating
effects of landscape change, such as habitat
loss, around sites targeted for conservation
(Cabeza, 2003; Moilanen, 2005b; Westphal et al.,
2007; I, V).
The use of spatial SHDMs has sometimes been
criticized as a quick fix for correcting for spatial
dependency in the observations, caused by a
failure to include the relevant variables as
covariates in the model (Lichstein et al., 2002;
Guisan & Thuiller, 2005). The results from
chapter II indicate that spatial SHDMs can be
used successfully for studying causes of
aggregation in species distribution patterns.
This can lead to a better understanding of
factors driving the spatial distribution of
species, and subsequently to biologically more
meaningful SHDMs. These findings can be
relevant for planning reserve networks with
appropriate levels of aggregation for species,
depending on the factors that drive their
distribution. Further work can encompass more
complex species-habitat relationships and the
study of time-series data to disentangle static
from dynamic exogenous variables.
Planning with multiple conservation options
per site is still in its infancy in reserve network
design and restoration planning literature,
despite related work in other fields such as
forest harvest scheduling (Hof & Raphael, 1993;
Hof et al., 1994; Bevers et al., 1995). This thesis
shows that the reserve selection problem can be
framed in a large context of single- and multiaction conservation planning problems, which
can incorporate several levels of spatial
complexity (Fig. 3). Reserve design algorithms
already found some applications in restoration
planning, but these were defined as singleaction problems (Newbold & Eadie, 2004;
Crossman & Bryan, 2006; Westphal et al., 2007).
This thesis provides novel straightforward
methods to solve multi-action planning
problems, a non-spatial method for multiple
species (III, IV), and an explicitly spatial
method for a single species or species with
similar requirements (V). Identifying ways to
explicitly account for the spatial requirements
of species with different habitat needs, in
management options (central area of Fig. 3), is a
challenging but common conservation planning
issue. Redirecting the methodology employed
in this thesis and theory from reserve network
design towards an integrated planning
framework, which allows for multiple actions
maintenance, provides ample opportunity for
future work.
This thesis would not have been written
without the help and support of many people.
First of all, I would like to thank Atte, for giving
me the opportunity to conduct my PhD
research in the Metapopulation Research Group
under his supervision. You have introduced me
to the world of reserve network design
methodology, thought me much about
computational issues and coding, and always
found time for answering questions. You gave
me the freedom to develop my own work,
regardless of when and where I was working,
which I very much appreciate. As a result, not
all of my projects have been your core business.
Nevertheless, you have always tried to keep me
on track and on schedule to make this PhD a
success, thank you!
I would like to thank all my co-authors for
being a great source of inspiration, knowledge
and reflection. Mar, although supervision has
not been your duty, you have always shown
interest in my work. You made me focus on the
questions and our discussions helped me
realise and understand what I was trying to do.
Thank you for taking the effort! Swimming in
the unfamiliar sea of spatial statistics, it took a
mathematician to retrieve the track. Otso, thank
you for your interest, for explaining me
complex things in simple terms (or silly
drawings), and for many constructive
comments. Katja, Juha and Mikko familiarised
me with mesic grassland communities, and
gave me the opportunity to consider the
strengths and limitations of my methods in a
practical context, which was a very educational
experience to me.
Though not having coauthored the work in this
thesis, other people have given an invaluable
contribution to my education. From them, I am
most grateful to Jane Elith and Brendan Wintle,
who taught me a great deal about habitat
modelling, without which I would not have
been able to complete this work. Thank you for
sharing your knowledge, and for being such
wonderful hosts during my visit to the
Environmenal Science Group in Melbourne. I
would also like to thank Henk Sierdsema and
Rogier Pouwels, for providing data for two of
my chapters and for answering numerous
questions on model parameters and result
interpretation. Risto Heikkinen and Kris
Rothley I would like to thank for their
encouraging reviewer statements, and Risto for
his efforts to comment the thesis in detail and
for providing me with food for thought.
These years would not have been so enjoyable,
if I would not have worked with great
colleagues, many of which became friends. The
MRG is an exceptional group to work in, and I
would like to thank Ilkka for providing us with
excellent research conditions. My sincere
thanks go to all MRG-members for your
willingness to discuss ideas, comment
manuscripts and presentations, and for the
pleasant daily atmosphere. I’ve also very much
enjoyed the reserve design reading club, thanks
Mar, Anni, Heini and Evgeniy! I would like to
thank Tapio, Anu, Mimma, Nina, Elina, Sami
and Tuuli for taking good care of the MRGsecretariat over the years - you have made my
life a lot easier. Of the very many social events
that have taken place I have warm memories,
thanks to Sofia, Ace, Itsuro, Luisa and all other
MRGideae for making these years such a great
The department has been a pleasant place to
work in, for there work a great number of fine
persons. The Wednesday’s coffee breaks, the
spring symposiums and research seminars
were inspiring events. Thanks to Andrés,
Hanna K, Katja B, Johan, Bob, Kata, Daniel,
Outi and many others for creating this
ambience. Since my office and the coffee room
were far apart, I oftentimes ran through the
corridor. I apologise to those who I have scared
when I almost bumped into them around the
corner! Veijo Kaitala, Ilkka Teräs and the
coordinators of the LUOVA graduate school:
Anna-Liisa Laine, Tomas Roslin and Heikki
Hirvonen, thank you for bridging the
bureaucratic hurdles.
A number of people have made my life in
Finland unforgettable, for which I would like to
express them my deepest gratitude. Tapio and
Mar have taken me on numerous trips berryand mushroom picking, camping, sailing,
skiing – experiences of the Finnish countryside
I would not have liked to have missed, I hope
that many more will follow. Mar & Tero, thank
you also for giving me a home during my visits.
Jonna has pulled me through the last difficult
stages of completing this work. I owe you for
the times you answered the phone even though
it made you miss your bus! Thanks for being a
friend, I sincerely miss your irony and face
expressions! Evgeniy, thank you for the pulla
we had, for the heaps of silly jokes and for
coping incredibly well with the angelcharacters in 5619, YM! Jenni, thanks an elephant
for the many board games where you let me
win(!), but most for your kindness, support and
for appreciating my baaad sense of humour!
Hanna, Elina, Annukka and Esko, thanks for
dragging me out of the office once in a while
and make me talk about something else than
work. Karin, Dieuwke, Elma, Karin, Marc,
Maarten and many others, thanks for your
support and friendship over the years, I’m
looking forward to catching up. Ton en Efi,
Mireille, Maurice, Fleur en Koen, Henny en
Hans, jullie wil ik heel hartelijk bedanken voor
jullie steun en betrokkenheid bij ons Finlandavontuur, super!
Mams en paps, bedankt voor de warmte, het
vertrouwen en de vrijheid die ik altijd van jullie
kreeg. Bedankt voor jullie steun en medeleven
tijdens mijn promotieonderzoek en de ontelbare
lieve kaartjes die op de deurmat vielen! Ik kan
me geen lievere ouders wensen. Niels and
Yvonne, thank you for your warm hospitality
that provides a wonderful home to us in family
holidays; those are days with a golden rim.
Lieve Emma, thank you being such as sunshine,
this thesis is for you. We have not been able to
visit as often as we would have liked, I hope we
can make up for that in the time to come.
Lieve Bart, your encouragement and support
have been indispensable during all this time.
Doing a PhD and living a large part of the time
in two countries requires trade-offs and
sacrifices. Thank you for your unconditional
love and sincerity, for the perseverance that I
sometimes lacked but needed so badly, and for
the many beautiful moments we’ve shared
these years. I’m looking forward to more!
My work has been generously funded by the
Academy of Finland (grant #202870 to A.
Moilanen), and I have been able to present my
work at several conferences with financial
support from to the Chancellor’s travel grant.
The University of Melbourne has kindly funded
my stay at the Environmental Science Group.
Adam, P. 1992. The end of conservation on the cheap.
National Parks Journal 36:19-22.
Andelman, S.J., and W.F. Fagan. 2000. Umbrellas and
flagships: Efficient conservation surrogates or
expensive mistakes? Proceedings of the National
Academy of Sciences of the United States of America
Andren, H., and P. Angelstam. 1988. Elevated
predation rates as an edge effect in habitat islands
- experimental evidence. Ecology 69:544-547.
Araújo, M.B., and P.H. Williams. 2000. Selecting areas
for species persistence using occurrence data.
Biological Conservation 96:331-345.
Arponen, A., R.K. Heikkinen, C.D. Thomas, and A.
Moilanen. 2005. The value of biodiversity in
weighting, and benefit functions. Conservation
Biology 19:2009-2014.
Arthur, J.L., M. Hachey, K. Sahr, M. Huso, and A.R.
Kiester. 1997. Finding all optimal solutions to the
reserve site selection problem: formulation and
computational analysis. Environmental and
Ecological Statistics V4:153-165.
Augustin, N.H., M.A. Mugglestone, and S.T.
Buckland. 1996. An autologistic model for the
spatial distribution of wildlife. Journal of Applied
Ecology 33:339-347.
Bakker, J.P., A.P. Grootjans, M. Hermy, and P.
Poschlod. 2000. How to define targets for
ecological restoration? Applied Vegetation Science
Besag, J.E. 1972. Nearest-neighbour systems and the
auto-logistic model for binary data. Journal of the
Royal Statistical Society B 34:75-83.
Bevers, M., J. Hof, B. Kent, and M.G. Raphael. 1995.
Sustainable forest management for optimizing
multispecies wildlife habitat: A coastal douglasfir example. Natural Resource Modeling 9:1-23.
Binzenhofer, B., B. Schröder, B. Strauss, R.
Biedermann, and J. Settele. 2005. Habitat models
and habitat connectivity analysis for butterflies
and burnet moths - The example of Zygaena
carniolica and Coenonympha arcania. Biological
Conservation 126:247-259.
Briers, R.A. 2002. Incorporating connectivity into
Conservation 103:77-83.
Brown, J.H., and A. Kodric-Brown. 1977. Turnover
rates in insular biogeography: effect of
immigration on extinction. Ecology 58:445-449.
Bruggeman, D.J., M.L. Jones, F. Lupi, and K.T.
Scribner. 2005. Landscape equivalency analysis:
Methodology for estimating spatially explicit
biodiversity credits. Environmental Management
Burgman, M.A., S. Ferson, and H.R. Akçakaya 1993.
Risk assessment in conservation biology.
Chapman and Hall, University press, Cambridge.
Cabeza, M. 2003. Habitat loss and connectivity of
reserve networks in probability approaches to
reserve design. Ecology Letters 6:665–672.
Cabeza, M., M.B. Araújo, R.J. Wilson, C.D. Thomas,
M.J.R. Cowley, and A. Moilanen. 2004a.
Combining probabilities of occurrence with
spatial reserve design. Journal of Applied Ecology
Cabeza, M., and A. Moilanen. 2001. Design of reserve
networks and the persistence of biodiversity.
Trends in Ecology & Evolution 16:242-248.
Cabeza, M., A. Moilanen, and H.P. Possingham.
2004b. Metapopulation dynamics and reserve
network design. In: I. Hanski, and O. Gaggiotti,
(Eds). Metapopulation ecology, genetics, and
evolution. Elsevier Academic Press, Amsterdam.
Camm, J.D., S. Polasky, A. Solow, and B. Csuti. 1996.
A note on optimal algorithms for reserve site
selection. Biological Conservation 78:353-355.
Chiarucci, A., F. Auria, V. De Dominicis, A. Lagana,
C. Perini, and E. Salerni. 2005. Using vascular
plants as a surrogate taxon to maximize fungal
species richness in reserve design. Conservation
Biology 19:1644-1652.
Church, R. 1996. Reserve selection as a maximal
covering location problem. Biological Conservation
Crossman, N.D., and B.A. Bryan. 2006. Systematic
programming. Biological Conservation 128:369-383.
Csuti, B., S. Polasky, P.H. Williams, R.L. Pressey, J.D.
Camm, M. Kershaw, A.R. Kiester, B. Downs, R.
Hamilton, M. Huso, and K. Sahr. 1997. A
comparison of reserve selection algorithms using
data on terrestrial vertebrates in Oregon.
Biological Conservation 80:83-97.
Debinski, D.M., and R.D. Holt. 2000. A survey and
overview of habitat fragmentation experiments.
Conservation Biology 14:342-355.
Dobson, A.P., A.D. Bradshaw, and A.J.M. Baker. 1997.
Hopes for the future: restoration ecology and
conservation biology. Science 277:515-522.
Donald, P.F., and A.D. Evans. 2006. Habitat
connectivity and matrix restoration: the wider
implications of agri-environment schemes. Journal
of Applied Ecology 43:209-218.
Dunford, R.W., T.C. Ginn, and W.H. Desvousges.
2004. The use of habitat equivalency analysis in
natural resource damage assessments. Ecological
Economics 48:49-70.
Elith, J., C.H. Graham, R.P. Anderson, M. Dudik, S.
Ferrier, A. Guisan, R.J. Hijmans, F. Huettmann,
J.R. Leathwick, A. Lehmann, J. Li, L.G. Lohmann,
B.A. Loiselle, G. Manion, C. Moritz, M.
Nakamura, Y. Nakazawa, J.M. Overton, A.T.
Peterson, S.J. Phillips, K. Richardson, R. ScachettiPereira, R.E. Schapire, J. Soberon, S. Williams,
M.S. Wisz, and N.E. Zimmermann. 2006. Novel
methods improve prediction of species'
distributions from occurrence data. Ecography
Faith, D.P., and P.A. Walker. 1996. How do indicator
groups provide information about the relative
biodiversity of different sets of areas?: On
hotspots, complementarity and pattern-based
approaches. Biodiversity Letters 3:18-25.
Ferrier, S., G.V.N. Powell, K.S. Richardson, G.
Manion, J.M. Overton, T.F. Allnutt, S.E. Cameron,
K. Mantle, N.D. Burgess, D.P. Faith, J.F.
Lamoreux, G. Kier, R.J. Hijmans, V.A. Funk, G.A.
Cassis, B.L. Fisher, P. Flemons, D. Lees, J.C.
Lovett, and R.S.A.R. Van Rompaey. 2004.
Mapping more of terrestrial biodiversity for
global conservation assessment. BioScience
Ferrier, S., G. Watson, J. Pearce, and M. Drielsma.
2002. Extended statistical approaches to
modelling spatial pattern in biodiversity in
northeast New South Wales. I. Species-level
modelling. Biodiversity and Conservation 11:2275–
Gaston, K.J., R.L. Pressey, and C.R. Margules. 2002.
biodiversity in the landscape and in protected
areas. Journal of Biosciences 27:361-384.
Gilbert, O.L., and P. Anderson 1998. Habitat creation
and repair. Oxford University Press, Oxford, UK.
Guisan, A., and W. Thuiller. 2005. Predicting species
distribution: offering more than simple habitat
models. Ecology Letters 8:993-1009.
Guisan, A., and N.E. Zimmermann. 2000. Predictive
habitat distribution models in ecology. Ecological
Modelling 135:147-186.
Haines, A.M., M.E. Tewes, L.L. Laack, W.E. Grant,
and J. Young. 2005. Evaluating recovery strategies
for an ocelot (Leopardus pardalis) population in the
United States. Biological Conservation 126:512-522.
Hanski, I. 1991. Single-species metapopulation
dynamics: concepts, models and observations.
Biological Journal of the Linnean Society 42:17-38.
Hanski, I. 1998. Metapopulation dynamics. Nature
Hanski, I. 1999. Metapopulation Ecology. Oxford
University Press, Oxford.
Hanski, I. 2005. The shrinking world: Ecological
consequences of habitat loss. International
Ecology Institute, Oldendorf/Luhe.
Harper, D.J., and J.T. Quigley. 2005. No net loss of
fish habitat: A review and analysis of habitat
Management 36:343-355.
Hilderbrand, R.H., A.C. Watts, and A.M. Randle.
2005. The myths of restoration ecology. Ecology
and Society 10.
Hof, J., M. Bevers, L. Joyce, and B. Kent. 1994. An
integer programming approach for spatially and
temporally optimizing wildlife populations.
Forest Science 40:177-191.
Hof, J., M. Bevers, D.W. Uresk, and G.L. Schenbeck.
2002. Optimizing habitat location for black-tailed
prairie dogs in southwestern South Dakota.
Ecological Modelling 147:11-21.
Hof, J., and M.G. Raphael. 1993. Some mathematical
programming approaches for optimizing timber
age-class distributions to meet multispecies
wildlife population objectives. Canadian Journal of
Forest Research 23:828-834.
Holzkämper, A., A. Lausch, and R. Seppelt. 2006.
Optimizing landscape configuration to enhance
habitat suitability for species with contrasting
habitat requirements. Ecological Modelling 198:277292.
Howard, P.C., P. Viskanic, T.R.B. Davenport, F.W.
Kigenyi, M. Baltzer, C.J. Dickinson, J.S. Lwanga,
R.A. Matthews, and A. Balmford. 1998.
Complementarity and the use of indicator groups
for reserve selection in Uganda. Nature 394:472475.
IUCN. 1993. Parks for Life - Report of the 4th World
Conference on Natural Parks and Protected
Areas. IUCN, Gland, Switzerland.
Kirkpatrick, J.B. 1983. An iterative method for
establishing priorities for the selection of nature
reserves: an example from Tasmania. Biological
Conservation 25:127-134.
Köhler, B., A. Gigon, P.J. Edwards, B. Krusi, R.
Langenauer, A. Luscher, and P. Ryser. 2005.
Changes in the species composition and
conservation value of limestone grasslands in
Northern Switzerland after 22 years of
contrasting managements. Perspectives in Plant
Ecology, Evolution and Systematics 7:51-67.
Legendre, P. 1993. Spatial autocorrelation: Trouble or
new paradigm? Ecology 74:1659-1673.
Leslie, H., M. Ruckelshaus, I.R. Ball, S. Andelman,
and H.P. Possingham. 2003. Using siting
algorithms in the design of marine reserve
networks. Ecological Applications 13:S185-S198.
Lichstein, J.W., T.R. Simons, S.A. Shriner, and K.E.
Franzreb. 2002. Spatial autocorrelation and
autoregressive models in ecology. Ecological
Monographs 72:445-463.
Lindenmayer, D.B., and M. Burgman 2005. Practical
Margules, C.R., and R.L. Pressey. 2000. Systematic
conservation planning. Nature 405:243-253.
Margules, C.R., R.L. Pressey, and P.H. Williams. 2002.
Representing biodiversity: data and procedures
for identifying priority areas for conservation.
Journal of Biosciences 27:309-326.
Margules, C.R., and J.L. Stein. 1989. Patterns in the
distributions of species and the selection of nature
reserves an example from eucalyptus forests in
south-eastern New South Wales Australia.
Biological Conservation 50:219-238.
McDonnell, M.D., H.P. Possingham, I.R. Ball, and
E.A. Cousins. 2002. Mathematical methods for
spatially cohesive reserve design. Environmental
Modeling and Assessment 7:107-114.
McLaughlin, J.F., J.J. Hellmann, C.L. Boggs, and P.R.
Ehrlich. 2002. Climate change hastens population
extinctions. Proceedings of the National Academy of
Sciences of the United States of America 99:60706074.
Millennium Ecosystem Assessment. 2005. Ecosystems
and human well-being: Biodiversity synthesis.
World Resources Institute, Washington, DC.
Moilanen, A. 2005a. Methods for reserve selection:
Interior point search. Biological conservation
Moilanen, A. 2005b. Reserve selection using nonlinear species distribution models. American
Naturalist 165:695-706.
Moilanen, A., and M. Cabeza. 2002. Single-species
dynamic site selection. Ecological Applications
Moilanen, A., and M. Cabeza. 2005. Variance and
uncertainty in the expected number of
occurrences in reserve selection. Conservation
Biology 19:1663-1667.
Moilanen, A., A.M.A. Franco, R.E. Early, R. Fox, B.A.
Wintle, and C.D. Thomas. 2005. Prioritizing
multiple-use landscapes for
Proceedings of the Royal Society of London. Series B,
Biological Sciences 272:1885-1891.
Moilanen, A., and M. Nieminen. 2002. Simple
connectivity measures in spatial ecology. Ecology
Moilanen, A., and B.A. Wintle. 2006. Uncertainty
analysis favours selection of spatially aggregated
reserve networks. Biological Conservation 129:427434.
Moilanen, A., and B.A. Wintle. 2007. The boundaryquality penalty: a quantitative method for
fragmentation in reserve selection. Conservation
Biology 21:355–364.
Moilanen, A., B.A. Wintle, J. Elith, and M.A.
Burgman. 2006. Uncertainty analysis for regionalscale reserve selection. Conservation Biology
Morris, R.K.A., I. Alonso, R.G. Jefferson, and K.J.
Kirby. 2006. The creation of compensatory habitat
- Can it secure sustainable development? Journal
for Nature Conservation 14:106-116.
Morris, W.F., and D.F. Doak 2003. Quantitative
conservation biology: theory and practice of
population viability analysis. Sinauer Associates,
Sunderland, Massachusetts.
Nalle, D.J., J.L. Arthur, and J. Sessions. 2002.
Designing compact and contiguous reserve
networks with a hybrid heuristic algorithm. Forest
Science 48:59-68.
Newbold, S., and J.M. Eadie. 2004. Using specieshabitat models to target conservation: A case
study with breeding mallards. Ecological
Applications 14:1384-1393.
Nicholson, E., and H.P. Possingham. 2006. Objectives
for multiple-species conservation planning.
Conservation Biology 20:871-881.
Nicholson, E., M.I. Westphal, K. Frank, W.A.
Rochester, R.L. Pressey, D.B. Lindenmayer, and
H.P. Possingham. 2006. A new method for
conservation planning for the persistence of
multiple species. Ecology Letters 9:1049-1060.
Nikolakaki, P. 2004. A GIS site-selection process for
habitat creation: estimating connectivity of
habitat patches. Landscape and Urban Planning
Önal, H. 2003. First-best, second-best, and heuristic
solutions in conservation reserve site selection.
Biological Conservation 115:55-62.
Önal, H., and R.A. Briers. 2002. Incorporating spatial
criteria in optimum reserve network selection.
Proceedings of the Royal Society of London. Series B,
Biological Sciences 269:2437-2441.
Opdam, P., J. Verboom, and R. Pouwels. 2003.
Landscape cohesion: an index for the
conservation potential of landscapes for
biodiversity. Landscape Ecology 18:113-126.
Pearson, D.L. 1994. Selecting indicator taxa for the
Philosophical Transactions of the Royal Society of
London Series B-Biological Sciences 345:75-79.
Pimm, S.L., and J.H. Lawton. 1998. Planning for
biodiversity. Science 279:2068–2069.
Pimm, S.L., G.J. Russell, J.L. Gittleman, and T.M.
Brooks. 1995. The future of biodiversity. Science
Polasky, S., J.D. Camm, A.R. Solow, B. Csuti, D.
White, and R. Ding. 2000. Choosing reserve
networks with incomplete species information.
Biological Conservation 94:1-10.
Possingham, H.P., I. Ball, and S. Andelman. 2000.
representative reserve network. Pages 291-306. In:
S. Ferson, and M. Burgman, (Eds). Quantitative
methods for conservation biology. Springer, New
Pöyry, J., S. Lindgren, J. Salminen, and M. Kuussaari.
2005. Responses of butterfly and moth species to
restored cattle grazing in semi-natural grasslands.
Biological Conservation 122:465-478.
Pressey, R.L. 1994. Ad hoc reservations: forward or
backward steps in developing representative
reserve systems? Conservation Biology 8:662-668.
Pressey, R.L., C.J. Humphries, C.R. Margules, R.I.
Vane-Wright, and P.H. Williams. 1993. Beyond
opportunism: Key principles for systematic
reserve selection. Trends in Ecology & Evolution
Pressey, R.L., H.P. Possingham, and J.R. Day. 1997.
Effectiveness of alternative heuristic algorithms
for identifying indicative minimum requirements
for conservation reserves. Biological Conservation
Pressey, R.L., H.P. Possingham, and C.R. Margules.
1996. Optimality in reserve selection algorithms:
When does it matter and how much? Biological
Conservation 76:259-267.
Pykälä, J. 2005. Plant species responses to cattle
grazing in mesic semi-natural grassland.
Agriculture, Ecosystems & Environment 108:109117.
Race, M.S., and M.S. Fonseca. 1996. Fixing
compensatory mitigation: what will it take?
Ecological Applications 6:94-101.
Roach, B., and W.W. Wade. 2006. Policy evaluation of
equivalency analysis. Ecological Economics 58:421433.
Rodrigues, A.S.L., and K.J. Gaston. 2001. How large
do reserve networks need to be? Ecology Letters
Rodrigues, A.S.L., and K.J. Gaston. 2002.
Optimisation in reserve selection procedures why not? Biological Conservation 107:123-129.
Root, K.V. 1998. Evaluating the effects of habitat
quality, connectivity, and catastrophes on a
threatened species. Ecological Applications 8:854865.
Saunders, D.A., R.J. Hobbs, and C.R. Margules. 1991.
Fragmentation: A Review. Conservation Biology
Schtickzelle, N., J. Choutt, P. Goffart, V. Fichefet, and
M. Baguette. 2005. Metapopulation dynamics and
conservation of the marsh fritillary butterfly:
Population viability analysis and management
options for a critically endangered species in
Western Europe. Biological Conservation 126:569581.
Schultz, C.B., and E.E. Crone. 2005. Patch size and
connectivity thresholds for butterfly habitat
restoration. Conservation Biology 19:887-896.
Smith, F., R. May, R. Pellew, T. Johnson, and K.
Walter. 1993. How much do we know about the
current extinction rate? Trends in Ecology &
Evolution 8:375-378.
Soulé, M.E., and M.A. Sanjayan. 1998. Ecology conservation targets: do they help? Science
Suding, K.N., K.L. Gross, and G.R. Houseman. 2004.
Alternative states and positive feedbacks in
restoration ecology. Trends in Ecology and
Evolution 19:46-53.
Svancara, L.K., R. Brannon, J.M. Scott, C.R. Groves,
R.F. Noss, and R.L. Pressey. 2005. Policy-driven
vs. evidence-based conservation: a review of
political targets and biological needs. BioScience
Tear, T.H., P. Kareiva, P.L. Angermeier, P. Comer, B.
Czech, R. Kautz, L. Landon, D. Mehlman, K.
Murphy, M. Ruckelshaus, J.M. Scott, and G.
Wilhere. 2005. How much is enough? The
recurrent problem of setting measurable
objectives in conservation. BioScience 55:835-849.
Ten Kate, K., J. Bishop, and R. Bayon. 2004.
Biodiversity offsets: Views, experience, and the
business case. IUCN, Gland, Switzerland and
Cambridge, UK and Insight Investment, London,
Thomas, C.D., A. Cameron, R.E. Green, M. Bakkenes,
L.J. Beaumont, Y.C. Collingham, B.F.N. Erasmus,
M.F. de Siqueira, A. Grainger, L. Hannah, L.
Hughes, B. Huntley, A.S. Van Jaarsveld, G.F.
Midgley, L. Miles, M.A. Ortega-Huerta, A.T.
Peterson, O.L. Phillips, and S.E. Williams. 2004.
Extinction risk from climate change. Nature
Environmental cluster analysis as a tool for
conservation sites. Ecological Applications 15:335345.
Underhill, L.G. 1994. Optimal and suboptimal reserve
selection algorithms. Biological Conservation 70:8587.
Vane-Wright, R.I., C.J. Humphries, and P.H.
Williams. 1991. What to protect? Systematics and
the agony of choice. Biological Conservation 55:235254.
Verboom, J., R. Foppen, P. Chardon, P. Opdam, and
P. Luttikhuizen. 2001. Introducing the key patch
approach for habitat networks with persistent
populations: an example for marshland birds.
Biological Conservation 100:89-101.
Vos, C.C., J. Verboom, P.F.M. Opdam, and C.J.F. Ter
Braak. 2001. Toward ecologically scaled
landscape indices. American Naturalist 157:24-41.
Westphal, M.I., S.A. Field, and H. Possingham. 2007.
Optimizing landscape configuration: A case
study of woodland birds in the Mount Lofty
Ranges, South Australia. Lanscape and Urban
Planning in press.
Westphal, M.I., and H.P. Possingham. 2003. Applying
a decision-theory framework to landscape
planning for biodiversity: Follow-up to Watson et
al. Conservation Biology 17:327-329.
Williams, J., C.S. ReVelle, and S.A. Levin. 2004. Using
mathematical optimization models to design
nature reserves. Frontiers in Ecology and the
Environment 2:98-105.
Williams, J.C., and S. Snyder. 2005. Restoring habitat
corridors in fragmented landscapes using
Environmental Modeling and Assessment 10:239-250.
Williams, P., D. Faith, L. Manne, W. Sechrest, and C.
Mapping the performance of surrogates for
biodiversity. Biological Conservation 128:253-264.
Williams, P.H., and M.B. Araújo. 2000. Using
probability of persistence to identify important
areas for biodiversity conservation. Proceedings of
the Royal Society of London Series B Biological
Sciences 267:1959-1966.
Williams, P.H., and M.B. Araújo. 2002. Apples,
oranges, and probabilities: Integrating multiple
factors into biodiversity conservation with
Assessment 7:139-151.
Williams, P.H., D. Gibbons, C.R. Margules, A. Rebelo,
C.J. Humphries, and R.L. Pressey. 1996. A
comparison of richness hotspots, rarity hotspots,
and complementary areas for conserving
diversity of British birds. Conservation Biology
Wilson, K., R.L. Pressey, A. Newton, M. Burgman, H.
Possingham, and C. Weston. 2005a. Measuring
and incorporating vulnerability into conservation
planning. Environmental Management 35:527-543.
Wilson, K.A., M.I. Westphal, H.P. Possingham, and J.
Elith. 2005b. Sensitivity of conservation planning
to different approaches to using predicted species
distribution data. Biological Conservation 122:99112.
Wintle, B.A., J. Elith, and J.M. Potts. 2005. Fauna
habitat modelling and mapping: A review and
case study in the Lower Hunter Central Coast
region of NSW. Austral Ecology 30:719-738.
Woodroffe, R., and J.R. Ginsberg. 1998. Edge effects
and the extinction of populations inside protected
areas. Science 280:2126-2128.
Young, T.P. 2000. Restoration ecology and
conservation biology. Biological Conservation
- -Yl