4 Principle 2 – Manage

connectivity
Vasilis Dakos, Allyson Quinlan, Jacopo A.
Baggio, Elena Bennett, Örjan Bodin and Shauna
BurnSilver

summary
Connectivity refers to the structure and strength with which
resources, species or social actors disperse, migrate or interact across
patches, habitats or social domains. Here we discuss how connectivity
may confer resilience to the supply of ecosystem services. High levels
of connectivity can facilitate recovery after a disturbance. At the same
time, highly connected systems increase the potential for distur-
bances to spread. Additionally, the structure characterizing how
system components are connected appears to play a role. Thus, the
effect of connectivity on the provision of ecosystem services is highly
context dependent. Despite increasing theoretical work that evaluates
how connectivity affects the resilience of social–ecological systems,
we still largely lack empirical studies that quantify these effects. We
discuss this disparity and suggest new areas for further research.

Principles for Building Resilience: Sustaining Ecosystem Services in Social–Ecological

Systems, eds R. Biggs, M. Schlüter and M. L. Schoon. Published by Cambridge University
Press. © Cambridge University Press 2015.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 81

4.1 introduction
Ecosystem services are integrated within systems in which social and
ecological components interact across multiple spatial and temporal
scales. For this reason, connectivity between these components can be
influential in conferring resilience to the production and/or supply of
specific sets of ecosystem services in response to disturbances. For
instance, isolated forest patches in a fragmented landscape may escape
fire (Peterson 2002), whereas highly connected agricultural landscapes
may suffer the spread of pests or diseases (Davis et al. 2008). Densely
connected social networks may facilitate governance of ecological
resources (Bodin and Prell 2011), although at the risk of reducing
diversity in management strategies that may potentially undermine
the resilience of the managed resources (Bodin and Norberg 2005;
McAllister et al. 2006). The implications of connectivity for the resi-
lience of ecosystem services are further complicated by the fact that
the above processes operate simultaneously at different scales. The
recent financial meltdown across the world economy and droughts in
major bread baskets, coupled with spikes in food prices, are clear
examples of how increasing global connectivity and novel intercon-
nections at different scales have implications for the resilience of such
complex systems (Biggs et al. 2011). In the following, we draw from
the growing body of theoretical work, and the few empirical studies
that explicitly test the relationship between connectivity and
resilience of ecosystem services, to explore how connectivity may
affect the resilience of the provision of ecosystem services to
disturbances in social–ecological systems (SES).

4.2 what do we mean by connectivity?

Connectivity refers to the way in which parts of an SES (i.e. entities
that have similar features such as species, landscape patches,
individuals, organizations and so forth) interact with each other
(i.e. exchange information, transfer material, transform energy, etc.).
For example, consider the case of forest patches and their connections

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 82 dakos, quinlan, baggio, bennett, bodin and burnsilver

in a landscape. The forested landscape is the system, the forest patches

are the parts of the system and their interactions dictate how easy it is
for an organism to move from one patch to another. Here, however, we
do not only consider connectivity in the context of a spatially explicit
landscape. In every system that can be conceptualized as a sum of
individual components, connectivity refers to the nature and the
strength of the interactions between these components. In that
sense, in any ecological or social ‘landscape’, connectivity is the way
by which (structure) and the extent to which (strength) resources,
species or social actors disperse, migrate or interact across patches,
habitats or social domains (Bodin and Prell 2011). While these inter-
actions are primarily mapped statically, they also change through
time (Raudsepp-Hearne et al. 2010). However, in this chapter, we do
not consider the consequences of temporal changes in connectivity,
but only focus on how the structure and strength of connectivity
influences the resilience of ecosystem services in coupled SES.
Alternatively, we can also think of connectivity from a network
perspective. In a network all individual components of a system are
nodes embedded in a web of connections that constitute the links
(Fig. 4.1). Examples of links are species interactions (like feeding
interactions in a food web), vegetation corridors across habitats or
communication channels between human communities (Table 4.1).
The way the links are distributed within an SES determines the struc-
ture of the SES. For instance, links may be present or absent between
components; they may be one-way interactions or mutual (reciprocal)
interactions. At the same time, some components/nodes may be
highly connected (i.e. have many links), while others may have few
connections (like an isolated patch of trees at the edge of a forest).
Links are also characterized by their strength, which refers to the
intensity with which nodes are connected or interact. Strength can
be determined by various factors, such as corridor quality among
habitats, preferences of a predator for specific prey, the visitation
rate of a pollinator insect to a plant or the frequency of interactions
between social actors.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 83

Random structure Nested structure Modular structure

Information interactions Spatial interactions Mutualistic interactions Trophic interactions
(social system) (ecological system) (food web)
Plants

Predators

Manager Habitat patch Producers Top predators

Pollinators

fig. 4.1 Toy representations of the architecture of connectivity in SES.

SES can be organized in random, nested or modular ways. In a random
network each node has on average the same number of links to other
nodes and no particular characteristic. Nested networks are usually bi-
partite (meaning that they are made of nodes that belong to a distinct
group). In nested networks, nodes interact with only a subset of nodes
in a hierarchical way. In a modular (or compartmentalized) network,
the nodes are organized in distinct compartments that are connected to
one another with very few links. Lines indicate interactions (edges/
links) between components (nodes), which can be of various sorts. For
example, in food webs, links between producers define competition
interactions, while links between predators and producers define
trophic links. Dashed lines indicate interactions across SES. The
thickness of the lines indicates the strength of interaction
(connectivity strength). Different shades and shapes correspond to
different types of system components (e.g. parts/actors/patches/
species).

4.3 how does connectivity enhance the

resilience of ecosystem services?
Connectivity in SES generally facilitates the flow of energy, material
or information necessary for the resilience of ecosystem services. In
particular, the strength and structure of connectivity may safeguard
ecosystem services against a disturbance either by facilitating recov-
ery or by constraining locally the spread of a disturbance (Nyström
et al. 2001).
The importance of connectivity for the recovery of disturbed
SES and thus for the maintenance of resilience of their ecosystem
services can be demonstrated in the recolonization of coral reefs.
The extent of reef recolonization is related to the degree of connectiv-
ity between remnant coral patches, which is determined by the

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 84 dakos, quinlan, baggio, bennett, bodin and burnsilver

Table 4.1 Examples of nodes and links for ecological and social
systems across different scales

Scale Ecological Social

Local Node Patch, species Individual,
household
Link Migration, Information,
energy (kcal) trust, food,
labour, money,
equipment
Intermediate Node Landscape Community,
(Regional) patches, organization (e.g.
populations, NGOs), firm,
trophic levels geographical
region
Link Migration, Information,
genetic material, opinion, ideas,
energy (kcal) trust, money,
expertise, food,
equipment,
labour
Global Node Ecosystems, Nation states,
biomes governance
entities,
stakeholder
groups
Link Migration, Rules, norms,
genetic material, decisions,
energy (kcal) information,
trust, finances

prevailing currents that allow coral recruitment between neighbour-

ing reefs (Treml et al. 2007; Mumby and Hastings 2008). Similarly, in
disturbance experiments of macrobenthic communities, recovery was
largely determined by the degree of connectivity across metacommu-
nities (Thrush et al. 2008). Closely situated habitats with no physical
barriers enhance the recolonization of nearby sites, safeguarding
against loss of species and ensuring the maintenance of their

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 85

functions. Thus, approaches of protecting, maintaining and restoring

connectivity underlie many conservation initiatives that focus on
enhancing the resilience of SES and the ecosystem services they pro-
duce, as in for instance the design of networks of marine protected
areas in the Great Barrier Reef off the coast of Australia (McCook et al.
2010). The basic mechanism behind these cases is that maintaining
connections to areas that serve as refuges can accelerate the restora-
tion of disturbed areas and their associated ecosystem services, as
these refuges provide a critical habitat that reduces larvae and juvenile
mortality and enhances recruitment in other less-resilient habitats
(Nyström et al. 2001).
Instead of facilitating recovery from a disturbance, connectivity
may also enhance the resilience of ecosystem services by acting as a
barrier to the spread of disturbances (like the spread of fire or of a
disease vector). This capacity is usually maximized in moderately
connected SES that are highly heterogeneous (P1 – Diversity). While
local disturbances can cause local regime shifts with local losses of
ecosystem services, limited connectivity reduces the possibility of
large-scale global effects (e.g. Bodin and Norberg 2007). In other
words, the potential loss of ecosystem services is locally contained
due to the existence of bottlenecks in the landscape. Generally, the
effect of connectivity on these two processes of recovery and con-
straint can be conceptualized by sink–source dynamics with compen-
sating effects (Dias 1996). Sources are parts of the system that produce
or maintain resources, whereas sinks are parts that do not. For exam-
ple, some reefs act as nursing grounds for fish and produce larvae
(sources), and other reefs are receiving larvae from the source reefs to
populate their own fish community (sinks). In marine systems, resi-
lient local systems may act as sources whereas non-resilient systems
act as sinks. Depending on the type and size of disturbance, the overall
resilience of systems that can maintain local fishing
communities (McCook et al. 2010) will be dependent on the strength
of the dispersal processes between local sink and source parts
(Nyström et al. 2008).

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 86 dakos, quinlan, baggio, bennett, bodin and burnsilver

Perhaps the most highlighted positive effect of connectivity on

the resilience of SES is that it facilitates the maintenance of biodiver-
sity in the landscape, which underlies the production of many ecosys-
tem services. The reason is that among well-linked habitat patches
local species extinctions may be compensated by the inflow of species
from their surroundings. This rescue effect (Hanski 1991) has been
demonstrated experimentally in moss microecosystems (Gonzalez
et al. 1998) and supports the idea of a spatial insurance hypothesis
(Loreau et al. 2003). According to this hypothesis spill-over effects
have been shown to be overall beneficial for the maintenance of
biodiversity (Brudvig et al. 2009) and for reducing the risk of extinc-
tion (Gilbert et al. 1998). Clearly, all these effects are a function of
connectivity in the landscape. As connectivity affects diversity (both
functional and genetic, P1 – Diversity), it will also indirectly have an
effect on the resilience of ecosystem services provided by a high
diversity. For instance, reduced connectivity caused by human-
induced fragmentation, like roads and dams, has a negative effect on
population viability (Fahrig and Rytwinski 2009) especially for large
mammal populations (Beier and Noss 1998). The Yellowstone-to-
Yukon project is an example of conservation planning that focuses
on reconnecting large habitat patches by establishing corridors to
minimize the effects of reduced genetic diversity in isolated large-
carnivore populations (Raimer and Ford 2005). Managers even mimic
connectivity in fragmented landscapes through additions of species or
individuals to decrease the risk of extinction of local populations. The
successful design of such schemes depends largely on conserving key-
stone patches in the landscape, creating new patches in the vicinity of
vulnerable ones or managing highly connected patches that all can
maximally contribute to the resilient provision of ecosystem services
(Janssen et al. 2006). However, the relationship between connectivity
and the maintenance of biodiversity is not straightforwardly linear.
Theoretical work suggests that a certain level of connectivity is
required in order to prevent extinctions, but an overly connected
system may reduce the probability of population survival when all

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 87

populations are experiencing the same strong exploitation practices

(Baggio et al. 2011; Salau et al. 2012).
Although a growing body of literature highlights the effects of
connectivity on resilience of SES, empirical evidence of the explicit
relationship of connectivity and the resilience of ecosystem services
remains limited (Mitchell et al. 2013). It has been shown that eco-
system functions that affect ecosystem productivity are influenced
by the strength of dispersal across habitats in the landscape (Staddon
et al. 2010), which infers a positive relationship of connectivity to
the resilience of a bundle of ecosystem services like provisioning and
regulating services. Another example is crop pollination by insects
that is tied to other ecosystem services such as food production and
habitat provision. It has been shown that connected patches in the
landscape have an effect on plant–animal interaction by increasing
pollination (Tewksbury et al. 2002) or changing patterns of butterfly
movements (Haddad et al. 2001). As pollination is strongly influ-
enced by the distance between plants that require pollination and
suitable habitats for their pollinators (Kearns et al. 1998; Ricketts
et al. 2008), it may be inferred that connectivity in such agroecosys-
tems will be important for the provision of pollination, food and
habitat services. Similarly, bundles of ecosystem services associated
with water flowing in streams and rivers within a watershed (e.g.
drinking-water provision, erosion control, water-quality regulation,
recreation opportunities, etc.) can be simultaneously affected by
changes in landscape connectivity of riparian buffering corridors.
How intact native vegetation remains along the shoreline of urban
streams affects their physical and biological condition, with less-
connected buffer zones having reduced erosion control leading to
poorer stream conditions (McBride and Booth 2005).
In human social networks, it has been argued that
connectivity can facilitate the resilience of ecosystem services
through enhanced governance opportunities. High levels of connec-
tivity between different social groups increase information-sharing
and help develop the trust and reciprocity necessary for collective

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 88 dakos, quinlan, baggio, bennett, bodin and burnsilver

action (P6 – Participation) (Brondizio et al. 2009). Certain actors can

serve as connectors to other actors or landscapes, bringing outside
perspectives and new ideas to local issues (P5 – Learning) (Bodin and
Crona 2009). Empirical studies in social networks demonstrate that
the strength of connectivity (e.g. frequency or duration of interactions)
may have different effects on the resilience of ecosystem services
depending on the aspirations and importance of the interacting social
actors. For example, it appears that actors are most likely to have
strong ties to actors with similar characteristics (Mcpherson et al.
2001). This homogeneity of characteristics can lead, for instance, to
a high connectivity of resource users with similar perspectives and
knowledge about the resources they exploit: the ‘who you know is
what you know’ phenomenon (Ruef 2002; Crona and Bodin 2006;
Bodin and Norberg 2007; Little and McDonald 2007). Increasing levels
of network connectivity across different social groups gives indivi-
duals opportunities for new information, and development of trust
and reciprocity necessary for collective action (Diani 2003). These
findings suggest that social network connectivity can facilitate resi-
lience of ecosystem services through enhanced governance (Bodin and
Crona 2009), while high levels of connectivity among actors with
similar types of knowledge or economic preferences can hinder col-
lective action or aid in cases of resource overexploitation of resources
(see Section 4.4).
Another relatively novel and largely unexplored feature of the way
connectivity affects the resilience of SES is the actual structure of SES.
Network theory suggests that non-random configurations (like modular
and nested structures) of food webs, mutualistic communities (seed-
dispersers–plants, plants–pollinators) or habitat patches have a positive
effect on their stability (Ash and Newth 2007; Galstyan and Cohen
2007; Bastolla et al. 2009) (Fig. 4.1). Modularity (or compartmentaliza-
tion) refers to the extent to which subsets of densely connected nodes
are loosely connected to other subsets of nodes, creating in essence
distinct compartments within a network (Fig. 4.1). Nestedness is the
degree to which specialist nodes (nodes with few links) interact with

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 89

subsets of generalist nodes (nodes with a lot of links) (Fig. 4.1).

Although progress to identify the effect of network architecture on
the resilience of SES is recent (Scheffer et al. 2012), it appears that
modular ecosystems, like lakes, are functionally organized in inde-
pendent modules. In that sense disturbances are difficult to spread
globally across modules or to cascade (Carpenter 2003), whereas
nested communities can sustain higher levels of disturbances but at
the cost of collapsing synchronously once a threshold is crossed
(Lever et al. 2014). These preliminary studies highlight that the
architecture of SES masks trade-offs between the provision of ecosys-
tem services and their resilience (Box 4.1).

4.4 under what conditions may resilience

of ecosystem services be compromised?
While connectivity can facilitate recovery or constrain the spread of a
disturbance, in some cases it can also compromise the resilience of
ecosystem services. Depending on the type and size of disturbance, in
strongly connected systems without compartmentalization (i.e. frag-
mented parts or local weakly connected parts) and with dense path-
ways between parts of the system, disturbances can propagate rapidly,
leading to widespread impacts on SES and associated ecosystem ser-
vices (Van Nes and Scheffer 2005; Ash and Newth 2007). Pest out-
breaks, disease epidemics, invasion of alien species, or even financial
crises, such as the global spread of the 2008 recession triggered by the
collapse of the US housing market, confirm the high risk of propaga-
tion of disturbances in strongly connected systems (Adger et al. 2009;
Biggs et al. 2011). In some ecosystems, such as the pine forests of
western North America, pest outbreak mechanisms rely on highly
connected patches (Raffa et al. 2008). An intricate set of factors and
conditions, all linked to high connectivity, are necessary to set in
motion a bark beetle infestation such as the pine beetle outbreak
that has been occurring across large expanses of forests in western
Canada and northwestern USA. Pheromone-triggered mass attacks on
trees by thousands of closely located beetles is facilitated by dense

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 90 dakos, quinlan, baggio, bennett, bodin and burnsilver

box 4.1 Social–ecological connectivity and

preservation of forests in southern
Madagascar
Taking both social and ecological connectivity into account when
analysing resilience is challenging from many perspectives. First, it
involves an explicit integration of disciplines that have different prac-
tices, epistemologies and methods. Second, there is the issue of defin-
ing relevant units of analysis (i.e. actors, ecological resources) and
appropriate types of links. Finally, linking patterns of connectivity to
resilience is not necessarily theoretically straightforward. All these
clearly pose a challenging task in deciphering the role of connectivity
in the resilience of ecosystem services. Here we illustrate how a com-
bined social and ecological network analysis can help to empirically
disentangle how patterns of social–ecological connectivity might influ-
ence resilience.
The study system is an agricultural landscape in southern
Madagascar that is interspersed with scattered forest patches of high
biodiversity value that have been remarkably well preserved in spite of
strong pressures on land and forest resources (Fig. 4.2) (Tengö et al.
2007; Bodin and Tengö 2012). The forest patches are protected by
taboos restricting access and use, and the patches generate essential
ecosystem services, such as micro-climate regulation and crop pollina-
tion (Bodin et al. 2006). Furthermore, the forests provide cultural
ecosystem services as they are important ancestral burial grounds,
sites for ceremonies and a symbol of the link between people and
land (Tengö et al. 2007).
The puzzle is whether an analysis of the patterns of social–ecological
connectivity can help explain why these forest patches have been
preserved in spite of an increasing demand for land. More specifically,
this example focuses on the control and use of ecosystem services
stemming from the forest patches. The forest patches are defined as
the nodes of the ecological part of the SES. These patches are geogra-
phically located within a village with approximately 9000 inhabitants.
In the social part of the SES, nodes are the social actors and are made up
of the six main clans in the village, alongside two additional clans
residing elsewhere but with a stake in the forests. The social links
between the clans were assessed through interviews with clan

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 91

box 4.1 Continued

(a)

(b)

fig. 4.2 The agricultural landscape (a) in Androy, southern

Madagascar, is interspersed with scattered forest patches (b) of high
biodiversity value that have been remarkably well preserved due to
various taboos restricting access and use (Tengö and von Heland 2011).
Photo credits: Maria Tengö.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 92 dakos, quinlan, baggio, bennett, bodin and burnsilver

box 4.1 Continued

authorities, and the ecological links between the forest patches were
assessed based on the potential for seed dispersal between patches
(Bodin and Tengö 2012). Finally, the social to ecological links were
defined based on access and use rights of the forest patches. The
complete social–ecological network describing the study system
consisted of 14 ecological and 8 social nodes (Fig. 4.3a).
A network analysis of the complete social–ecological network
would involve the estimation of the average number of links, the
distribution of links among nodes or the relative compartmentaliza-
tion for each part of the SES network. In this example, however, it was
more sensible to use a micro-scale approach. This is because a simpli-
fied network might be easier to link, at least theoretically, to some
governance challenges of interest that have implications for the resi-
lience of ecosystem services. In particular, by analysing how often a set
of three micro-scale social–ecological networks (also called ‘motifs’)
occurred in the complete network, it was found that shared forest
access generally implied social connectivity (Fig. 4.3b). In other
words, if two clans were utilizing one and the same forest patch, they
tended to also be socially connected. This means that resource sharing
(i.e. competition) is often accompanied with social connectivity.
According to common-pool resource theory (Ostrom 1990), this con-
figuration of social–ecological connectivity may increase the potential
for negotiating and regulating resource use in a sustainable way. Thus,
the relatively high frequency of this presumably favourable motif in
the larger network may provide a possible explanation of why the
forest patches have been preserved in this otherwise heavily exploited
agricultural landscape.
There is another overrepresented motif in the network (Fig. 4.3b).
That is the motif where the clans share two forest patches that are also
connected. Thus, highly interconnected clusters of clans and forest
patches seem to occur relatively frequently. The third most common
motif is the one of symmetric pair-wise coupling of social and ecolo-
gical nodes (Fig. 4.3b). The occurrence of such motif patterns indicates
that (1) socially connected clans tend to share sets of patches that are
ecologically connected and (2) unconnected patches are divided
between unconnected clans. Overall, this implies a positive alignment
of social and ecological patterns of connectivity, or social–ecological

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 93

box 4.1 Continued

a
b Motif 1
Clan

Forest
patch

Motif 2

Motif 3

0 0.5 1 km

fig. 4.3 The social–ecological network of an agricultural landscape

in southern Madagascar. Dark grey areas represent the 14 forest
patches (ecological nodes of the SES), and the circled black dots are
the 8 clans that are maintaining the forest and constitute the social
nodes of the SES. The white lines are the interrelationships between
the clans, while the solid lines are the dispersal rates of seeds between
the forest patches based on a measure of distance. The dasked lines
depict the relationship (ownership, management) of each clan to each
forest patch. (b) The micro-scale social–ecological networks (motifs).
Motif 1 shows shared forest patch access combined with social
connectivity between clans. In motif 2, complete connectivity
between clans and forest patches exists, and in motif 3 there is a pair-
wise coupling of the clans and the forest patches.

‘fit’ that generally suggests increased capacity to manage ecosystems in

a sustainable way.
In summary, this example demonstrates how, by analysing connec-
tivity patterns of SES, one may partly explain the successful govern-
ance of biodiversity-rich forest patches. Such analysis of specific
patterns of connectivity of SES may highlight mechanisms to enhance
the resilience of ecosystem services.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 94 dakos, quinlan, baggio, bennett, bodin and burnsilver

host-tree stands that have become structurally similar due to

environmental factors such as drought or temperature (Raffa et al.
2008). In this case, the resilience of ecosystem services, such as
wood resource production and pest regulation, is negatively impacted
by high connectivity.
High levels of connectivity between social actors can also be
detrimental to the resilience of ecosystem services through a
different mechanism. In this case, strong links may lead to synchro-
nized behaviour that translates into intense unsustainable resource
extraction or strong norm barriers for giving up unsustainable prac-
tices (Bodin and Prell 2011). For example, Satake et al. (2007) found
that in dense social networks with preferences for immediate gains
over long-term resilience, information about a change in market
price for timber can spread quickly and result in deforestation as
multiple actors take advantage of market conditions to cut and sell
their timber at the same time. Moreover, highly connected networks
may limit social learning and reduce the capacity to find optimal
solutions and reduce the ability for novel experimentation (P5 –
Learning). For example, modelling studies show that when homo-
genization of norms occurs, explorative ability drops, leading to a
lock-in situation in which actors believe themselves to be doing well
while they are actually driving their managed ecosystems towards
unsustainable pathways (Bodin and Norberg 2005). In the case of
SES, these results suggest that overly connected networks can lead
to homogenization of strategies when imitation and social influence
are strong. Lack of diversity may lead to maladaptation: instead
of people adapting to the changing environment, the effect is that
the environment itself must adapt to human norms (Levinthal
1997), with negative consequences for the resilience of
ecosystem services.
Another way that connectivity can jeopardize the resilience of
ecosystem services is by weakening or disrupting links across differ-
ent compartments. For instance destruction of breeding or refuge
grounds of species with diverse life cycles disrupts their recruitment

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 95

and can make SES less resilient. Mumby and Hastings (2008) have
shown that shallow coral reefs become less resilient due to the
destruction of mangroves that serve as breeding grounds for parrotfish
larvae that play a crucial part in the functioning of the reef. In this
case, one of the services provided by mangroves, i.e. provision of
parrotfish breeding habitat, is disrupted. Such types of interdependen-
cies are commonly found across ecological and social domains at
multiple scales (Brondizio et al. 2009). It has been suggested that in
recent years, due to the increased interlocking of systems like markets
and global food distribution networks, local disruptions or scarcities
had disproportional impacts on the resilience of SES on distance scales
that have been unknown so far (Biggs et al. 2011).
In highly modular (compartmentalized) systems, resilience of
ecosystem services may be jeopardized if some components become
overly important compared to others, in the sense that the whole
system relies heavily on some individual parts (Strogatz 2001).
Removal of important components, such as keystone species or highly
connected patches, may trigger cascading waves of extinctions
(P1 – Diversity). For example, models of Madagascar's dry-forest
dynamics suggest that rapid declines in pollination services could
occur if small forest patches are removed from the landscape, owing
to their impacts on the spatial configuration of the remaining forest
area (Bodin et al. 2006). If subgroups that actively use certain ecosys-
tem services are not engaged in the management of those ecosystem
services, critical knowledge of systems' functioning and monitoring
can be missed (P5 – Learning) (Gelcich et al. 2006), and there may be a
reduced potential for collective action (P6 – Participation).

4.5 how can the principle of connectivity

be operationalized and applied?
Operationalizing connectivity to enhance the resilient provision of
ecosystem services is clearly an ambitious goal. As with all principles,
operationalization is inevitably context-dependent. Different SES will
have different potentials for intervention, while the resilience of

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 96 dakos, quinlan, baggio, bennett, bodin and burnsilver

individual ecosystem services will be variably affected. Thus, system-

specific properties will determine how connectivity should be mana-
ged (e.g. by modulating strength, manipulating structure). Here we
attempt to postulate some general principles for operationalizing and
managing connectivity, although further research is needed to sub-
stantiate these principles. We discuss these gaps in the next section.

• Map connectivity. In order to understand the effect of connectivity on the

resilience of ecosystem services, the first step is to clearly map the
relevant parts of the SES, their scale, interactions and the currency of
connectivity (see Table 4.1 and Box 4.1). Based on this information,
visualization and network analysis tools can assist to map the structure
(random, nested, modular), the units flowing across links (information,
animals, energy, resources) and the strength (weak, moderate, strong) of
connectivity in the SES. Such mapping can help determine the overall
level of connectivity to develop hypotheses of the relative effect of
connectivity on the resilience of ecosystem services.
• Identify important elements/interactions. To guide interventions and
optimize the effect of connectivity on the resilience of ecosystem services,
network tools can help to identify keystone, highly connected nodes or
isolated patches in the SES. This helps to identify vulnerable and resilient parts
in the SES, and highlight groups of actors that are overly connected. In parallel,
it is important to characterize these elements in terms of their diversity (P1 –
Diversity) as the combination of their structure and uniqueness can determine
their relative importance. By doing so, one can rank nodes based on their
potential effect on the resilience of ecosystem services and can explore the
potential for managing structure and/or strength of connectivity in the SES
through particular nodes or interactions.
• Restore connectivity. The restoration of connectivity involves the creation
of nodes, or conservation of keystone nodes in the SES depending on the
ecosystem service that needs to be enhanced and the type of disturbance
against which one wants to protect. Alternatively, one can also maintain
or strengthen connections to more resilience nodes that serve as refuge
nodes. For example, habitat loss caused by the conversion of forest to crop or
pasture land or the disruption of blue corridors that convey fish, nutrients
and sediment when dams are built in rivers are a few examples that
demonstrate how disrupting connectivity can jeopardize the provision and

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 97

resilience of ecosystem services these systems provide. Restoring

connectivity in such cases can directly restore ecosystem functioning and
the supply of ecosystem services. For example, the Yellowstone-to-
Yukon (Y2Y) project in North America, mentioned above, focuses on
reconnecting large patches of wildlife habitat by restoring corridors
between patches. The approach of Y2Y primarily involves purchasing
land or enacting land trusts in large, intact watersheds, connecting habitats
to ensure species and genetic diversity (Baldwin et al. 2012). Similarly,
the Montérégie Connection project in southern Quebec, Canada, is about
connecting forests and people to make the landscape, and its provision
of ecosystem services, more resilient to climate change and other global
and regional changes (http://www.monteregieconnection.com). Here,
the focus is on a much smaller scale; the science primarily addresses
how the size and connectedness of forest patches is likely to
influence the provision of ecosystem services across the region
(Mitchell et al. 2013).
• Optimize current connectivity patterns. In other cases, it may be more
useful to manage current connectivity patterns to minimize the effect of a
disturbance on the resilience of ecosystem services and to contain the risk
of a systemic failure. This can be achieved by disrupting the connections
of extremely vulnerable components from the rest of the SES. Alternatively,
the goal might be to increase modularity in the structure of the SES in
order to create compartments that can act as bottlenecks for containing a
disturbance. The loss of electricity across the eastern USA and Canada in
2003, affecting an estimated 50 million people, is an example of a
network where local failures could not be contained and the system
experienced a systemic collapse. In this case, the high degree of
connectivity and lack of modularity of the power grid was largely
responsible for the extensive blackout (Andersson et al. 2005). Thus, in
some cases, connectivity may need to be reduced or otherwise changed
structurally to increase the resilience of the system. Nonetheless,
determining and achieving the appropriate structure and degree of strength
remains a challenge. This especially applies when analyses take social and
ecological connectivity into account simultaneously. Some attempts have
recently been taken in this direction, although much work remains to be
done for fully integrated social–ecological connectivity analyses (Box 4.1)
(Bodin and Tengö 2012; Schoon et al. 2014).

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 98 dakos, quinlan, baggio, bennett, bodin and burnsilver

4.6 key research and application gaps

Connectivity in an SES can facilitate recovery after a disturbance in an
ecological landscape or the development of trust necessary for collec-
tive action in social systems. However, at the same time, highly con-
nected systems increase the potential for disturbances to spread and
enhance the risk of homogenization of knowledge, which can lead to
suboptimal management. The relationship between connectivity and
the resilience of SES or the ecosystem service they provide is far from
simple. Despite the bulk of theoretical work that evaluates how con-
nectivity affects the resilience of SES, we are largely lacking empirical
studies. New work (e.g. Mitchell et al. 2013; Ziter et al. 2013) that
summarize the effect of connectivity on the provision of ecosystem
services are emerging, but still much remains to be understood regard-
ing the impact of connectivity on the resilience of ecosystem services.
This gap in our knowledge stems mostly from the difficulty in
identifying and measuring connectivity. The suggestions for opera-
tionalizing connectivity outlined in Section 4.5 are not easy to gen-
eralize. This is due to the large number of currencies employed for
quantifying connectivity in SES (e.g. flow of energy, resources,
information, interaction strengths, species movements) that make
it challenging to quantify and compare the strength of connectivity
across different SES. Moreover, difficulties arise in defining the
boundaries and/or agents in network representations of SES
(Bodin and Prell 2011). Most network theory has focused on
well-defined subsets of ecological interactions, like feeding relation-
ships in food webs or mutualistic interactions in plant–pollinator
communities, but the integration of multiple interaction types is
still largely unexplored (Fontaine et al. 2011; Mougi and Kondoh
2012). The consequences of integrating such a diversity of
interactions for the resilience of ecological systems are unclear
(Scheffer et al. 2012) and at the same time are difficult to grasp
conceptually. Furthermore, when one adds social network
dynamics, the task of disentangling the effects of connectivity on

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 99

the resilience of ecosystem services becomes even more challenging.

So far there is little understanding on the effect of integrating both
types of interactions (social and ecological) on the resilience of SES.
This challenge is aggravated by the fact that connectivity is not a
constant property, as the strength and structure of links may vary
over time.
Equally challenging is to know how to practically manage con-
nectivity to enhance the resilience of ecosystem services. Despite some
general statements about how to operationalize the level and type of
connectivity to improve the resilience of ecosystem services, future
research needs to focus on clear applications of connectivity manage-
ment based on a theoretical understanding of the impacts of connectiv-
ity. As the tools and methods for estimating ecosystem services
improve, initiatives that aim to quantify the impact of restoring land-
scape connectivity on ecosystem services will be better able to demon-
strate how specific actions translate into increased resilience of
ecosystem services.
Despite these challenges, recent advances in network tools for
mapping and analysing complex relationships in SES, as well as rapid
developments of quantifying ecosystem services, create new opportu-
nities for testing how managing the connectivity of SES may increase
the resilience of the ecosystem services they provide. At the same
time, increasing access to information such as remote-sensing data,
combined with (a) novel representations of social–ecological interac-
tions (Box 4.1), (b) individual-based modelling for simulating SES and
(c) theoretical advancements in understanding the effect of architec-
ture on the dynamics of complex networks may all offer new insights
into the relationship between connectivity and resilience of ecosys-
tem services.

references
Adger, W. N., Eakin, H., Winkels, A. (2009). Nested and teleconnected vulnerabil-
ities to environmental change. Frontiers in Ecology and the Environment, 7,
150–157.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 100 dakos, quinlan, baggio, bennett, bodin and burnsilver

Andersson, G., Donalek, P., Farmer, R. et al. (2005). Causes of the 2003 major grid
blackouts in North America and Europe, and recommended means to improve
system dynamic performance. IEEE Transactions on Power Systems, 20,
1922–1928.
Ash, J. and Newth, D. (2007). Optimizing complex networks for resilience against
cascading failure. Physica A: Statistical Mechanics and its Applications, 380,
673–683.
Baggio, J. A., Salau, K., Janssen, M. A., Schoon, M. L., Bodin, Ö. (2011). Landscape
connectivity and predator–prey population dynamics. Landscape Ecology, 26,
33–45.
Baldwin, R.F., Reed, S.E., McRae, B.H., Theobald, D.M., Sutherland, R.W. (2012).
Connectivity restoration in large landscapes: modeling landscape condition and
ecological flows. Ecological Restoration, 30, 274–279.
Bastolla, U., Fortuna, M.A., Pascual-Garcia, A. et al. (2009). The architecture of
mutualistic networks minimizes competition and increases biodiversity.
Nature, 458, 1018–1020.
Beier, P. and Noss, R. (1998). Do habitat corridors provide connectivity?
Conservation Biology, 12, 1241–1254.
Biggs, D., Biggs, R., Dakos, V., Scholes, R. J., Schoon, M. L. (2011). Are we entering
an era of concatenated global crises? Ecology and Society, 16, 27.
Bodin, Ö. and Crona, B. I. (2009). The role of social networks in natural resource
governance: what relational patterns make a difference? Global Environmental
Change, 19, 366–374.
Bodin, Ö. and Norberg, J. (2005). Information network topologies for enhanced local
adaptive management. Environmental Management, 35, 175–193.
Bodin, Ö. and Norberg, J. (2007). A network approach for analyzing
spatially structured populations in fragmented landscape. Landscape Ecology,
22, 31–44.
Bodin, Ö. and Prell, C., eds. (2011). Social Networks and Natural Resource
Management: Uncovering the Social Fabric of Environmental Governance.
Cambridge: Cambridge University Press.
Bodin, Ö. and Tengö, M. (2012). Disentangling intangible social–ecological sys-
tems. Global Environmental Change, 22, 430–439.
Bodin, Ö., Tengö, M., Norman, A., Lundberg, J., Elmqvist, T. (2006). The value of
small size: loss of forest patches and ecological thresholds in southern
Madagascar. Ecological Applications, 16, 440–451.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 101

Brondizio, E. S., Ostrom, E., Young, O. R. (2009). Connectivity and the governance
of multilevel social–ecological systems: the role of social capital. Annual
Review of Environment and Resources, 34, 253–278.
Brudvig, L. A., Damschen, E. I., Tewksbury, J. J., Haddad, N. M., Levey, D. J. (2009).
Landscape connectivity promotes plant biodiversity spillover into non-target
habitats. Proceedings of the National Academy of Sciences USA, 106,
9328–9332.
Carpenter, S. R. (2003). Regime Shifts in Lake Ecosystems: Pattern and Variation.
Oldendorf/Luhe, Germany: International Ecology Institute.
Crona, B. and Bodin, Ö. (2006). What you know is who you know? Communication
patterns among resource users as a prerequisite for co-management. Ecology
and Society, 11, 7.
Davis, S., Trapman, P., Leirs, H., Begon, M., Heesterbeek, J. A. (2008). The abun-
dance threshold for plague as a critical percolation phenomenon. Nature, 454,
634–637.
Diani, M. (2003). Introduction: social movements, contentious actions, and social
networks: from metaphor to substance. In Social Movements and Networks:
Users Without a Subscription are Not Able to See the Full Content. Social
Movements and Networks: Relational Approaches to Collective Action.
Oxford: Oxford University Press, pp. 1–20.
Dias, P. C. (1996). Sources and sinks in population biology. Trends in Ecology and
Evolution, 11, 326–330.
Fahrig, L. and Rytwinski, T. (2009). Effects of roads on animal abundance: an
empirical review and synthesis. Ecology and Society, 14, 21.
Fontaine, C., Guimarães, P. R Jr, Kéfi, S. et al. (2011). The ecological and evolu-
tionary implications of merging different types of networks. Ecology Letters, 14,
1170–1181.
Galstyan, A. and Cohen, P. (2007). Cascading dynamics in modular networks.
Physical Review E, 75, 36109.
Gelcich, S., Edwards-Jones, G., Kaiser, M.J., Castilla, J. C. (2006). Co-management
policy can reduce resilience in traditionally managed marine ecosystems.
Ecosystems, 9, 951–966.
Gilbert, F., Gonzalez, A., Evans-Freke, I. (1998). Corridors maintain species rich-
ness in the fragmented landscapes of a microecosystem. Proceedings of the
Royal Society of London B, 265, 577–582.
Gonzalez, A., Lawton, J. H., Gilbert, F. S., Blackburn, T. M., Evans-Freke, I. (1998).
Metapopulation dynamics, abundance, and distribution in a microecosystem.
Science, 281, 2045–2047.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 102 dakos, quinlan, baggio, bennett, bodin and burnsilver

Haddad, N. M., Tilman, D., Haarstad, J., Ritchie, M., Knops, J. M. (2001).
Contrasting effects of plant richness and composition on insect communities:
a field experiment. American Naturalist, 158, 17–35.
Hanski, I. (1991). Single-species metapopulation dynamics: concepts, models and
observations. Biological Journal of the Linnean Society, 42, 17–38.
Janssen, M. A., Bodin, Ö., Anderies, J. M. et al. (2006). Toward a network perspective
of the study of resilience in social–ecological systems. Ecology and Society, 11, 15.
Kearns, C. A., Inouye, D. W., Waser, N. M. (1998). Endangered mutualisms: the
conservation of plant–pollinator interactions. Annual Review of Ecology and
Systematics, 29, 83–112.
Lever, J. J., van Nes, E. H., Scheffer, M., Bascompte, J. (2014). The sudden collapse of
pollinator communities. Ecology Letters, 17, 350–359.
Levinthal, D. A. (1997). Adaptation on rugged landscapes. Management Science, 43,
934–950.
Little, L. R. and McDonald, A. D. (2007). Simulations of agents in social networks
harvesting a resource. Ecological Modelling, 204, 379–386.
Loreau, M., Mouquet, N., Gonzalez, A. (2003). Biodiversity as spatial insurance in
heterogeneous landscapes. Proceedings of the National Academy of Sciences
USA, 100, 12765–12770.
McAllister, R. R., Gordon, I. J., Janssen, M. A., Abel, N. (2006). Pastoralists'
responses to variation of rangeland resources in time and space. Ecological
Applications, 16, 572–583.
McBride, M. and Booth, D. B. (2005). Urban impacts on physical stream condition:
effects of spatial scale, connectivity, and longitudinal trends. Journal of the
American Water Resources Association, 98195, 565–580.
McCook, L. J., Ayling, T., Cappo, M. et al. (2010). Adaptive management of the
Great Barrier Reef: a globally significant demonstration of the benefits of net-
works of marine reserves. Proceedings of the National Academy of Sciences
USA, 107, 18278–18285.
Mcpherson, M., Smith-lovin, L., Cook, J. M. (2001). Birds of a feather: homophily in
social networks. Annual Reviw of Sociology, 27, 415–444.
Mitchell, M. G. E., Bennett, E. M., Gonzalez, A. (2013). Linking landscape connec-
tivity and ecosystem service provision: current knowledge and research gaps.
Ecosystems, 16, 894–908.
Mougi, A. and Kondoh, M. (2012). Diversity of interaction types and ecological
community stability. Science, 337, 349–351.
Mumby, P. J. and Hastings, A. (2008). The impact of ecosystem connectivity on
coral reef resilience. Journal of Applied Ecology, 45, 854–862.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 principle 2 – manage connectivity 103

Nyström, M. and Folke, C. (2001). Spatial resilience of coral reefs. Ecosystems, 4,

406–417.
Nyström, M., Graham, N. A. J., Lokrantz, J., Norström, A. V. (2008). Capturing the
cornerstones of coral reef resilience: linking theory to practice. Coral Reefs, 27,
795–809.
Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for
Collective Action. Cambridge: Cambridge University Press.
Peterson, G. D. (2002). Contagious disturbance, ecological memory, and the emer-
gence of landscape pattern. Ecosystems, 5, 329–338.
Raffa, K.F., Aukema, B. H., Brentz, B. J. et al. (2008). Cross-scale drivers of natural
disturbances prone to anthropogenic amplification: the dynamics of bark beetle
eruptions. BioScience, 58, 501.
Raimer, F. and Ford, T. (2005). Yellowstone to Yukon (Y2Y) – one of the largest
international wildlife corridors. Ecological Perspectives for Science and
Society, 14, 182–185.
Raudsepp-Hearne, C., Peterson, G. D., Bennett, E. M. (2010). Ecosystem service
bundles for analyzing tradeoffs in diverse landscapes. Proceedings of the
National Academy of Sciences USA, 107, 5242–5247.
Ricketts, T. H., Regetz, J., Steffan-Dewenter, I. et al. (2008). Landscape effects on
crop pollination services: are there general patterns? Ecology Letters, 11,
499–515.
Ruef, M. (2002). Strong ties, weak ties and islands: structural and cultural predictors
of organizational innovation. Industrial and Corporate Change, 11, 427–449.
Salau, K., Schoon, M. L., Baggio, J. A., Janssen, M. A. (2012). Varying effects of
connectivity and dispersal on interacting species dynamics. Ecological
Modelling, 242, 81–91.
Satake, A., Leslie, H. M., Iwasa, Y., Levin, S. A. (2007). Coupled ecological–social
dynamics in a forested landscape: spatial interactions and information flow.
Journal of Theoretical Biology, 246, 695–707.
Scheffer, M., Carpenter, S. R., Lenton, T. M. et al. (2012). Anticipating critical
transitions. Science, 338, 344–348.
Schoon, M.L., Baggio, J.A., Salau, K., Janssen, M. (2014). Insights for managers from
modeling species interactions across multiple scales in an idealized landscapes.
Environmental Modelling and Software, 55, 53–59.
Staddon, P., Lindo, Z., Crittenden, P. D., Gilbert, F., Gonzalez, A. (2010).
Connectivity, non-random extinction and ecosystem function in experimental
metacommunities. Ecology Letters, 13, 543–552.
Strogatz, S. H. (2001). Exploring complex networks. Nature, 410, 268–276.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005
 104 dakos, quinlan, baggio, bennett, bodin and burnsilver

Tengö, M. and von Heland, J. (2011). Adaptive capacity of local indigenous institu-
tions – the case of the taboo forests in southern Madagascar. In Adapting
Institutions Governance, Complexity, and Social–Ecological Resilience.
Boyd, E. and Folke, C., ed. Cambridge: Cambridge University Press, pp. 37–74.
Tengö, M., Johansson, K., Rakotondrasoa, F. et al. (2007). Taboos and forest govern-
ance: informal protection of hot spot dry forest in southern Madagascar.
AMBIO, 36, 683–691.
Tewksbury, J. J., Levey, D. J., Haddad, N. M. et al. (2002). Corridors affect plants,
animals, and their interactions in fragmented landscapes. Proceedings of the
National Academy of Sciences USA, 99, 12923–12936.
Thrush, S. F., Halliday, J., Hewitt, J. E., Lohrer, A. M. (2008). The effects of habitat
loss, fragmentation, and community homogenization on resilience in estuaries.
Ecological Applications, 18, 12–21.
Treml, E. A., Halpin, P. N., Urban, D. L., Pratson, L. F. (2007). Modeling population
connectivity by ocean currents, a graph-theoretic approach for marine conserva-
tion. Landscape Ecology, 23, 19–36.
Van Nes, E. H. and Scheffer, M. (2005). Implications of spatial heterogeneity for
regime shifts in ecosystems. Ecology, 86, 1797–1807.
Ziter, C., Bennett, E.M., Gonzalez, A. (2013). Functional diversity and management
mediate aboveground carbon stocks in small forest fragments. Ecosphere, 4, 1–21.

Downloaded from https://www.cambridge.org/core. North Carolina State University, on 08 Nov 2018 at 12:32:37, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781316014240.005