Journal for Nature Conservation 22 (2014) 504–515

Contents lists available at ScienceDirect

Journal for Nature Conservation

journal homepage: www.elsevier.de/jnc

Modelling land use changes for landscape connectivity: The role of

plantation forestry and highways
Soledad Nogués ∗ , Alejandro Cabarga-Varona
GEURBAN Research Group, Dpt. of Geography, Urban and Regional Planning, School of Civil Engineering, University of Cantabria, Av. Los Castros, 44,
39005 Santander, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Landscape connectivity is a key aspect for the maintenance of biodiversity and ecosystem viability.
Received 4 December 2013 Nowadays, the competition between economic development and nature conservation is intense. In most
Received in revised form 1 August 2014 territories natural vegetation is being replaced by exotic tree plantations, which have a better perfor-
Accepted 1 August 2014
mance in terms of timber productivity but often, a lower ecological value. We evaluated potential natural
forest connectivity improvement in the Cantabria region (Northern Spain) through two main actions:
Keywords:
protection of environmentally valuable forest areas, and reforestation with indigenous species of those
Forest dwelling-species
patches of exotic plantation trees with a particularly important role for the connectivity of the forest
Graph theory matrix
IIC index
network. We established a variety of scenarios to calculate least cost paths, considering the presence
PC index or absence of plantation forestry and highways to examine connectivity. Then, we applied two habitat
Reforestation availability indices (integral index of connectivity and probability of connectivity) attending to differ-
Road network ent dispersal distances. Our analyses show a great potential for improving connectivity using plantation
forests in the natural forest network, and a dramatic impact of the highway in the north–south con-
nectivity of the study area. Based on these results, we identified those patches of plantation forest and
natural forest that are more important for the maintenance of overall landscape connectivity, and propose
their protection or conversion through reforestation. The final proposed network constitutes a larger and
better connected natural forested landscape than the existing one.
© 2014 Elsevier GmbH. All rights reserved.

Introduction altered due primarily to human actions. Indeed, mosaics of land-

scape patches can have a natural origin, but are mainly caused by
Connectivity can be synthesised as the degree in which anthropogenic activities, such as plantation forestry, agricultural
landscape facilitates or impedes species movements and other eco- intensification, road network, or urbanization (Di Giulio et al. 2009;
logical flows (Rubio & Saura, 2012; Taylor et al. 1993), and is one of Fu et al. 2010; Li et al. 2010; Liu et al. 2008, 2011, 2014b; Szabó et al.
the key factors maintaining the ecological functions of forests (Liu 2012; Yu et al. 2012). In highly populated areas, landscape frag-
et al. 2014a). When forests are fragmented, the pattern of spatially mentation limits species survival to the existence of connectivity
structured habitats is modified and the movement of dispersing between spatially separated populations (Fischer & Lindenmayer,
individuals may be constrained, hampering biodiversity conser- 2007; Kramer-Schadt et al. 2004).
vation (Laita et al. 2010; Saura & Torné, 2009). This is in fact the Deforestation is one of the activities inflicted by humans which
case for many forest and ground-dwelling species (Pascual-Hortal has the most influenced landscape fragmentation (e.g. Fearnside,
& Saura, 2006, 2008), whose habitat availability has been severely 2005; Harper et al. 2007; Skole & Tucker, 1993). Clearance has
sometimes been aimed at the substitution of native tree species
with fast-growing ones, often exotic. These new species contribute
significantly to the economic growth of many regions as a result
of the derived benefits of timber exploitation and the functions
∗ Corresponding author at: Dpt. Geografía, Urbanismo y Ordenación del Territorio,
that they provide as forest habitats (see e.g. Brockerhoff et al. 2003,
E.T.S. Ingenieros de Caminos, Canales y Puertos, Universidad de Cantabria, Av. Los
2005; Carnus et al. 2006; Humphrey et al. 2000). However, they
Castros 44, 39005 Santander, Spain. Tel.: +34 942201780; fax: +34 942201703.
E-mail addresses: [email protected] (S. Nogués), may also induce substantial changes in natural ecosystems and
[email protected], [email protected] (A. Cabarga-Varona). habitat structure (Calviño-Cancela et al. 2012; Fabiao et al. 2002;
URL: http://grupos.unican.es/geurban/ (A. Cabarga-Varona). Poore & Fries, 1985) and it is usually true that natural forests offer

http://dx.doi.org/10.1016/j.jnc.2014.08.004
1617-1381/© 2014 Elsevier GmbH. All rights reserved.
 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515 505

better quality habitats for native forest species than plantation ones identified as pertinent for forest network connectivity, with native
(Brockerhoff et al. 2008). species.
Plantation forests can enhance indigenous biodiversity by
improving connectivity between natural forest remnants (see e.g.
Material and methods
Brockerhoff et al. 2008 and references therein). In fact, some
plantation forests can be assimilated to natural ones when their
Study area
management is sustainable, and is allowed to acquire “old-growth”
conditions (Humphrey, 2005). On the other hand, when planta-
The county of Trasmiera, is a ca. 600 km2 historic and natu-
tion forestry is very intensive, with short rotations and high timber
ral district located in the region of Cantabria (N of Spain), in the
productivity performance, plantation patches with a key role in net-
coastal plain (Fig. 1). Trasmiera is limited to the north by the Atlantic
work connectivity are eventually transformed into autochthonous
Ocean, and is constrained to the east and west by two bays, respec-
vegetation forest. Thus, previously existing plantations can bene-
tively, limiting its connection with the rest of the territory to the
fit the restoration of natural forests, accelerating natural recovery
southern stretch, which is bordered by the A-8 national highway
by modifying physical and biological conditions positively (Lamb
(Fig. 1). In the past, at least 50% of the area was occupied by Quercus
et al. 2005 and references therein). These management actions
robur, Quercus ilex and bushes, while crops and pastures occu-
however, need to be contemplated in the global perspective of
pied a smaller area than present (García-Martino, 1862). Various
the modified landscapes where they are planned and not as iso-
preindustrial activities contributed to the deforestation of the area,
lated actions, since specific topographical or man-made features
mainly from the 18th to the 19th century (ship construction, can-
may modify their results substantially. Large highways or high
non and conventional foundries, mining, etc.), but the cellulose
speed railway infrastructures, for instance, may pose important
industry was the decisive landscape change agent in the 20th cen-
barriers to connectivity notwithstanding the effect of plantation
tury, replacing native forest and agricultural-grassland/pastures
patches, and their effect should be included in the connectiv-
patches mainly with Eucalyptus sp., but also with plantation pines
ity analyses before a reforestation strategy is approved or put in
(P. radiata and P. pinaster) in a smaller proportion and in specific
motion.
areas. In relation to this industry, the reforestation policy through-
Indeed, these infrastructures are a major element in the frag-
out the 20th century did not include reforestation with native
mentation of natural habitats, and pose particular problems to
species and focused on the planting of these high-yield species.
forest dwelling fauna species which often encounter insurmount-
The populations inhabiting natural forest patches are thus
able barriers in particularly broad or busy highways (Forman
currently severely segregated from southern populations by the
& Alexander, 1998; Liu et al. 2014b; Spellerberg, 1998). Forest
highway, which is devoid of fauna crossings and has only three
patches are often severely segregated by roads limiting con-
main viaducts under which communication with the other side of
nectivity between forest-dwelling species. Their impacts can be
the highway is facilitated.
intensified by the flow of traffic (Langevelde et al. 2009), espe-
cially in border areas (see e.g. Forman & Deblinger, 2000), where
noise, pollution, luminosity, waste, etc. are more intense (Forman Approach to connectivity analysis
et al. 2002). On the other hand, roads have also been found to
attract specific faunal species (Dean & Milton, 2003; Dodd et al. Landscape connectivity is currently viewed either structurally,
2007), some being able to disperse through landscapes with major where connectivity is entirely based on landscape pattern (e.g. size,
roads or highways (Blanco et al. 2005; Waller & Servheen, 2005). shape, etc.), or functionally, where behavioural responses to land-
The impact of these infrastructures will also vary depending on scape elements are considered together with the spatial structure of
the sensitivity of the affected habitats and landscapes, and the the landscape (Tischendorf & Fahrig, 2000; Ziólkowska et al. 2012).
tolerance and adaptability of distinct animal and plant species liv- The assessment of functional connectivity usually requires more
ing in the area (Forman et al. 2002; Geneletti, 2006; Rytwinski & resources, since it entails monitoring species movements in the
Fahrig, 2013). For instance, the dispersal ability of organisms across area. Conversely, structural connectivity can be evaluated directly
changing landscapes, which is a fundamental issue for long-term from the spatial configuration of the landscape mosaic (Taylor et al.
biodiversity conservation (Fahrig, 2007), is a clear limiting fac- 2006).
tor when considering species abilities to move between preferred Among the different methods proposed to evaluate connec-
habitat patches, or surmount specific obstacles such as roads or tivity (Ziólkowska et al. 2012), there are two complementary
highways. approaches (Rubio et al. 2012), that have significantly contributed
Reforestation schemes aimed at guaranteeing connectivity to an improved and operational forest landscape connectivity anal-
between forested areas must take all these factors into account and ysis: graph theory and habitat availability (reachability) metrics
specifically analyse the implications that the presence of a partic- (for an overview see Galpern et al. 2011; Urban & Keitt, 2001).
ular barrier, such as a prominent highway or railway, may have for Graph theory models the relationships among nodes of a network.
achieving the planned connectivity objectives. It can be applied to “patch-based graphs”, where the nodes are the
We evaluated the potential connectivity improvement of the habitat patches and the links suggest the potential connections for
forest habitats network in the county of Trasmiera (Cantabria, the species (Galpern et al. 2011). Conversely, habitat availability
Spain) for medium and large sized mammals. To achieve this, we metrics consider a patch as a space where connectivity occurs (the
analysed the connectivity of natural and plantation forests, under larger the patch, the larger the connected area), integrating habi-
various land use scenarios and considered the specific influence tat patch area (or other patch attributes) and connections between
of a major infrastructure barrier: a highway, by applying the inte- different patches in a single measure (Pascual-Hortal & Saura,
gral index of connectivity (IIC), and the probability of connectivity 2006). Graph theory has given rise to many connectivity measures,
(PC) index. We identified those patches of plantation forest with some of them specifically designed for the evaluation of landscape
the best connectivity performance to include them in the natural connectivity (Laita et al. 2011). We used the integral index of con-
forest network and analysed their restoration possibilities through nectivity (IIC) and the probability of connectivity (PC), which are
reforestation. Finally, we proposed two main courses of action: (1) based both on graph theory and on the habitat availability concept
protection of environmentally valuable forest areas; and (2) refor- (Pascual-Hortal & Saura, 2006, 2008; Saura & Pascual-Hortal, 2007).
estation of relevant patches of exotic plantation trees previously Both indices provide a reasonably detailed picture of potential
506 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515

Fig. 1. Location of the study area in the Cantabria region (Spain). Dark green areas: natural forest; light green areas: plantation forests; solid red line: highway (A-8); circles:
three viaducts considered as current non-specific wildlife crossings; river symbol: estuaries. (For interpretation of the references to colour in this figure legend, the reader
is referred to the web version of this article.)

connectivity, whilst having relatively modest data requirements IIC and PC indices (Pascual-Hortal & Saura, 2006; Saura & Pascual-
(Saura & Torné, 2009). Graph-theory has the added value of per- Hortal, 2007). Node importance is given by the expressions:
mitting the analysis of patch linkage not only to assess overall
functional connectivity (Watts & Handley, 2010), but also at various IIC − IICi
dIIC (%) = · 100
distance thresholds. This allows addressing landscape connectiv- IIC
ity from the species’ perspective (Laita et al. 2010). The indices are
computed as (for more details see e.g. Pascual-Hortal & Saura, 2008; PC − PCi
dPC (%) = · 100
Saura & Pascual-Hortal, 2007): PC
n n ai ·aj where IIC or PC correspond to the overall index value calculated for
i=1 j=1 1+nlij the landscape (considering all the habitat patches), and IICi or PCi is
IIC =
A2L the overall index value after removing patch i from the landscape.
These indices were applied to the various scenarios proposed
where ai is the descriptive variable of each forest habitat area (node) (Table 1), which were then used to achieve the objectives raised
and nlij is the number of links in the shortest path (topological initially. To unravel which natural and plantation forest patches
distance) between nodes i and j. play an important role in the connectivity pattern we considered
n n the influence of highway and plantation forest. We focused our
i=1
a a p∗
j=1 i j ij
PC = 2
AL Table 1
Scenarios computed.
where ai and aj are the areas of habitat patches i and j, respectively,
Scenarios: composition of With highway Without highway
and AL is the total landscape area (that of the region under study,
resistance layer
comprising both habitat and non-habitat patches). It is defined as
the maximum product probability of all possible paths between Natural forest network 1a 1b
Natural + plantation forest network 2a 2b
patches i and j. For PC we defined a dispersal probability of 0.5
when the threshold distance equals 2000, 5000, 10,000, 15,000 and Scenarios 1a and 1b show the actual landscape connectivity pattern in the study
area considering only natural forests, and they allow us to identify the differences
25,000 m.
between the two resistance layers, with and without highway. Scenarios 2a and 2b
We calculated dIIC and dPC metrics, that is, the individual incorporate plantation forestry patches with lower resistance values than scenarios
importance of every single forest habitat patch in terms of per- 1a and 1b, equaling their value to that of natural patches. In this case, we also include
centage variation in the total degree of connectivity as given by the landscape resistance layers with and without highways.
 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515 507

connectivity analysis on terrestrial species, particularly those with Table 2

Resistance values for the movement resistance layer.
greater dispersal abilities, such as medium and large sized mam-
mals. Since these groups include species with differing dispersal Land use/cover Resistance
abilities, and following the abundant literature on this issue (e.g.
1a 1b 2a 2b
Decout et al. 2012; Fu et al. 2010; García-Feced et al. 2011;
Natural forest 0 0 0 0
Gurrutxaga et al. 2011; Laita et al. 2010, 2011; Rubio & Saura, 2012;
Bushes 4 4 4 4
Saura et al. 2011a; Szabó et al. 2012), we considered a variety of dis- Plantation forestry 15 15 0 0
persal or threshold distances when calculating the indices (2000, Agriculture-livestock 8 8 8 8
5000, 10,000, 15,000 and 25,000 m), in order to identify the most Water bodies 100,000 100,000 100,000 100,000
useful patches for differing dispersal abilities. Artificial (urban areas) 10,000 10,000 10,000 10,000

We considered those forested areas mapped in the vector ver- Linear road features added to the layer
sion of the Spanish Forest Map (SFM) scale 1:50,000 (MMA, 2002) Highways 100,000 8 100,000 8
National road 200 200 200 200
and the maps from the National Geographical Institute (IGN, 2006).
1st category regional road 50 50 50 50
We distinguished two networks: patches with predominantly 2nd category regional road 40 40 40 40
native species (forest habitat areas composed of autochthonous 3rd category regional road 30 30 30 30
species), and patches with predominantly exotic species (plan-
Slopes-values added to the rest of land covers
tation forestry, Eucalyptus globulus and Pinus spp.) or any other 0–15◦ 1 1 1 1
forested patches with some exotic plantation trees. 15–30◦ 2 2 2 2
◦
In order to prepare the data for further steps in our analysis, we 30–45 3 3 3 3
considered separately mixed patches (with native and exotic vege- 45–60◦ 4 4 4 4
60–75◦ 5 5 5 5
tation) and those having only one tree species. This information, in >75◦ 500 500 500 500
conjunction with the values of the indices, was used to determine
Note that highways and plantation forestry covers change resistance values depend-
the proposed replacement, restoration or protection actions to be
ing on the selected scenarios of analysis. To erase the highway from the layer
taken. With this objective we maintained information regarding we assigned a standard friction value (the most common land cover in the area:
the secondary and tertiary species present on each patch associated agriculture-livestock) so as to simulate a pre-implantation friction of this infrastruc-
with each layer. ture.
We tried to include as much spatial information (patches) as
possible to obtain precise results regarding those patches which
are relevant at landscape level. Thus, and considering time and spatial barriers that hinder individuals’ movements, such as roads
processing capability limitations, we decided to include all patches and water bodies. The road network was given higher resistance
≥1 ha. This size included patches which can serve as temporary values than the remaining land uses/covers. Roads received differ-
shelters or stepping stones for many species, as well as being a part ent resistance values attending to their category (i.e. size, building
of the surrounding landscape. characteristics and traffic volume). Water bodies corresponded to
estuary or bay areas which for most animals are difficult to cross.
Calculation of movement resistance layer and least-cost path Resistance values for highway and urban areas were assigned
distances to ensure that the constructed least-cost paths would not cross
the barriers unless no other possibility of movement existed (see
We used a resistance map or matrix, which is usually based on a e.g. Decout et al. 2012; Ziólkowska et al. 2012). That is, despite
grid composed by cells, each of them with values of friction to travel. the existence of other high resistance areas (e.g. a large pixel area
We proposed general resistance values (Table 2), based on values with an accumulated resistance value of 500), urban areas or high-
previously accepted in the literature and on our knowledge of the ways would not be crossed. It is worth noting that the A-8 highway
topography and particular barriers present in the study area. These has no specific wildlife crossings in the study area, the only areas
resistance values also simulate the friction that particular land cov- offering some permeability to the forest fauna being three viaducts
ers, especially those which are more inhospitable, such as urban which coincide with marginal river corridors. These structures are
areas and large infrastructures pose for the majority of wild fauna. between 100 and 200 m wide, and offer higher permeability than
The lowest resistance values were assigned to habitat patches other crossings specifically designed for vehicles, which makes
which may function either as habitat or stepping-stones. Resistance them the most likely wildlife passes currently crossing the A-8
in the matrix (e.g. for pastures and crop fields) was considered highway. Thus, we edited the original highways polygon, consider-
higher than in natural forests patches, which represent the node ing three stretches of least resistance in the sections corresponding
network in route calculation and have resistance values equal to 0, with the three viaducts (>15 m) present in the study area (Fig. 2). To
but lower than in hindering barriers (Pereira et al. 2011). We con- calculate the least-cost paths we built a resistance layer (landscape
sidered bushes to have a low resistance, given their presumable matrix) which took into account the land uses/covers included in
role as stepping-stone patches. On the other hand, we assigned a the SFM. We added information to the forest map layer, overlapping
higher resistance to plantation forestry, which in our study area the road network. To prevent the creation of discontinuities in nar-
is characterised by a typical fast-wood plantation monoculture, row roads (i.e. <10 m) due to the high level of detail of the matrix
impoverishing the understory (mainly composed of brackens and (5 m pixel), we applied a buffer to roads doubling their width. In
brambles, Pteridium aquilinum and Rubus ulmifolius, respectively). this way, we could assess more realistically the alterations to land-
In addition, the rotations are very short (around 15 years), and the scape connectivity caused by these infrastructures. In making this
area is totally cleared, generating an increase in the risk of fires, ero- edition, we modified the spatial information of the forest map,
sion, sudden edge effects, invasive species (e.g. Cortaderia selloana), intersecting the original polygons and generating new patches,
etc. These characteristics usually hamper animal displacements, also of forest, which are the real ones when the road network is
but may benefit particular species, which may be more adapted considered.
to move through densely vegetated areas. Agriculture–livestock Finally, we added resistance values in relation to the existing
patches are mainly comprised of pastures with few obstacles, but slope gradients, but attributed high (determinant) values under
higher human presence, which causes an increase in resistance val- very high slope conditions. To include all the land uses/covers we
ues with respect to bush areas. The highest values were assigned to edited the layer in vector format and then converted it into a raster
508 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515

Fig. 2. Images of Trasmiera showing the various landscapes/arrangements of vegetation (A and B) Typical plantations, composed of Eucalyptus globulus. (C) Natural forests
(Quercus robur, Quercus ilex, Corylus avellana, etc.) found near the coast and segregated from inner forest networks. (D) One of the three viaducts considered in the analysis.

grid with a 5 m cell size. This level of detail was the same as that results were obtained through the Conefors’ Sensinode 2.6 option
of the digital elevation model used to calculate the slopes (which “there are nodes to add” considering all 476 patches in the network.
together with land uses/covers and roads constituted the friction If the delta value (for IIC and PC) of the patch was higher than the
layer). mean value, the patch was selected for its good connectivity. We
only selected those patches in which both the dIIC and dPC values
were above the mean.
Node importance analysis and improvement of the landscape
connectivity network with reforestation
Data processing
We first analysed the connectivity status in the natural forest
network with and without highway (scenarios 1a and 1b), taking Management of the spatial data (vector layers edition, vector to
into account different threshold distances to identify the adequacy raster conversions and development of the resistance surface) was
of the system for the various fauna species. In this sense, we pro- performed using ArcGis 10.0 GIS software (ESRI® ArcMapTM 10.0).
posed to protect natural nodes since they are the basis of the natural Least cost path distances were obtained using the ArcGis extension
network. Linkage Mapper 1.0.2 (McRae & Kavanagh, 2011). Calculation of
Secondly, by incorporating plantation forestry patches we could the connectivity indices and its delta values was conducted with
identify the most valuable ones in terms of importance for connec- the ArcGis application Conefor Sensinode 2.6 (Saura & Torné, 2009)
tivity (dIIC and dPC), as well as the changes experimented in the which has already been applied in numerous studies. The results
connectivity importance of the natural network (scenarios 2a and from Linkage Mapper can be easily used as an input for the Conefor
2b) with their addition. We analysed plantation patch composi- Sensinode (connection file) by converting the least-cost path data
tion in more detail (considering secondary and tertiary dominant table into a text file with the requirements of Conefor.
species) to propose restoration actions in those cases in which
autochthonous species were still present. In the case of patches Results
exclusively composed by plantation species, we considered their
substitution by natural arborous species, in order to maintain the Currents patterns of connectivity of the Trasmiera natural forests
prevailing forested land use.
To do this, we considered scenario 2a (with highway, since it is The natural forest network was composed of 187 patches occu-
the real situation) and we took into account the IIC and PC impor- pying 13% (71.83 km2 ) of the total study area (592 km2 ). Differences
tance values. Patch selection for restoration and substitution was in the area of natural network and in the final results in comparison
based on the mean value of the dIIC and dPC results of the 289 plan- with Table 3 are due to the fact that in the connectivity analysis we
tation forestry patches at all dispersal distances (e.g. mean value considered only patches >1 ha. There are a few patches of highly val-
n = 289, for d = 2000, 5000, etc.). Thus, we obtained a synthetic value ued climactic potential vegetation dominated by Quercus ilex (62%,
capturing those patches which most increased connectivity. The relictic in the region), Quercus robur (20%) and Fagus sylvatica (5%).
 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515 509

Table 3 Table 5
Area occupied by land uses/covers and forest species in Trasmiera. dIIC and dPC values for 1a and 1b scenarios.

Land use/cover Area (Ha) Area (Ha) % % 2000 5000 10,000 15,000 25,000

Roads – 1383 – 2 dIIC

Highways 185 – 13 – Max 27.56 24.88 18.79 16.32 13.65
Scenario 1a (with
National roads 115 – 8 – Mean 0.900 0.906 0.991 1.037 1.032
highways)
Autonomic road 1 222 – 16 – Sum 169.25 170.49 186.35 194.97 194.17
Autonomic road 2 441 – 32 – Scenario 1b Max 27.54 24.76 18.61 16.06 14.01
Autonomic road 3 420 – 30 – (without Mean 0.900 0.907 0.997 1.042 1.175
Agriculture-livestock – 31,273 – 53 highways) Sum 169.23 170.69 187.58 195.96 221.06
Water bodies – 334 – 1
dPC
Artificial – 1966 – 3
Max 36.16 27.50 21.37 18.34 15.58
Natural forest – 7713 – 13 Scenario 1a (with
Mean 1.148 1.088 1.095 1.108 1.114
Quercus ilex 4789 – 62 – highways)
Sum 215.88 204.60 205.94 208.47 209.56
Quercus robur 1537 – 20 –
Scenario 1b Max 36.02 27.33 21.17 18.09 15.27
Fagus sylvatica 387 – 5 –
(without Mean 1.148 1.088 1.097 1.111 1.118
Castanea sativa 76 – 1 –
highways) Sum 215.92 204.71 206.28 209.05 210.32
Other species 924 – 12 –
Plantation forestry – 12,292 – 21
Eucalyptus globulus 11,768 – 96 –
Pinus radiata, Pinus pinaster 524 – 4 – values, Table 5). This effect was already apparent from 10,000 m of
Wetlands – 1383 – 2 dispersal distance and especially from 15,000 to 25,000 m.
Bushes – 2906 – 5
The location of important patches at the various dispersal
Total – 59,250 – 100 distances measured by the dIIC and dPC indicators, shows how
Source: MMA (2002) and IGN (2006). centrally located patches have better values at larger dispersal dis-
Area data of tree species have been calculated taking into account the dominant tances.
species in each patch, so that the surface of some secondary species may be slightly The most significant effects of highway presence were noted
higher.
at 25,000 m of dispersal distance for the dIIC, smaller differences
occurring in dPC values (Table 5). It should be highlighted that
However, they are surrounded by anthropogenic modified land, a small group of patches (between 10 and 15% of the total) con-
the current landscape matrix being a mosaic of livestock and agri- centrated from 60 to 80% of the total dIIC and dPC sum, and
cultural lands, plantation forestry, urban areas, roads, and natural thus accounted for most of the connectivity importance under
forests (Table 3, IGN, 2006; MMA, 2002). the various scenarios considered. In addition, it is interesting
Patches of native forest differed notably in their dIIC and dPC how maximum dIIC and dPC values (Table 5) were higher in the
values, indicating the importance of particular patches for main- highway-included scenario at all threshold distances (with the
taining connectivity. Larger patches which were located optimally exemption of dIIC d = 25,000).
from a connectivity perspective obtained the best results. Table 4
and Figs. 3 and 4, show how connections decreased whenever the Adding plantation forestry patches to natural network
highway was considered, and that the larger the threshold dis-
tance, the greater the differences as regard connections between We added 289 plantation forestry patches (122.83 km2 , 21% of
the patches. the study area) to the native network, increasing the number of
This is in close connection with the resistance that the landscape patches from 187 to 476. The number of links between scenarios
matrix represents for the longer least-cost paths, which only have 2a and 2b decreased the larger the threshold distance, especially
three ways of crossing the A-8 (highway) in the study area. The for d = 10,000, d = 15,000 and d = 25,000 (Table 6). The average value
highway-included matrix generated an average distance increase of the least-cost scenarios for a total of 113,050 connections was
of 10% (17,156 m) on the least-cost paths. Conversely, the matrix 115,090 m (with highway) and 101,497 m without the highway. In
without highways, ranged from 88,682 to 71,526 m, considering this case we also identified the same differences between the two
the 17,392 connections established between all the 187 patches of landscape matrix scenarios, with a decreasing number of connec-
forest habitat areas. tions in the scenario including highway.
This indicates that the highway generates greater travel needs Dispersal distances had a strong effect on our connectivity esti-
for animals. We observed a cluster of routes or connections which mates, with generalised increases in dIIC and dPC values the larger
created interconnected patch groups. At dispersal distances shorter the distance. However, from 5000 m onwards, the identified incre-
than 5000 m, the natural forest network of the study area fol- ment rate was lower (Fig. 5).
lows the pattern of a habitat which is concentrated in a few areas, There was a group of patches with diverse locations that con-
with relatively few isolated patches in the rest of the territory (see centrated between 2 and 10% of the global connectivity importance
Figs. 3 and 4). These groups of patches were located mainly in the in both the dIIC and the dPC results. On the other hand, about 10
southern zone of the study area, although they also appeared in the patches concentrated almost 40% of the total delta sum, a lower
northwestern part. At larger dispersal distances the highway influ- concentration than in scenarios 1a and 1b. We found that all the
enced the connectivity importance of the patches negatively (mean best connectivity providers were patches located near landscape

Table 4 Table 6
Number of links at various dispersal distances in scenarios 1a and 1b. Number of links at various dispersal distances in scenarios 2a and 2b.

Dispersal distance (m) No links 1a No links 1b Dispersal distance # links 2a # links 2b

2000 708 730 2000 1213 1229

5000 1877 1985 5000 3201 3293
10,000 4632 5102 10,000 7664 8273
15,000 7436 8595 15,000 13,622 15,233
25,000 12,080 14,140 25,000 26,831 33,931
510 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515

Fig. 3. Results of the dIIC index in scenarios 1a and 1b. The importance of the patches is classified into five levels based on IIC as very high importance, high importance,
medium importance, low importance, and very low importance (see e.g. Fu et al., 2010), using the natural breaks data classification of ArcMap.

units that had been previously identified (scenarios 1a and 1b) as connectivity was improved (Table 7). We selected plantation
critical. forestry patches with the best results of landscape connectivity,
We observed that by adding the new 289 plantation forestry based on the criteria specified in the node importance analysis
patches to the natural forest network (composed of 187), (delta values higher than the mean and species composition). The
 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515 511

Fig. 4. Results of dPC index in scenarios 1a and 1b. The importance of the patches is classified into five levels based on PC as very high importance, high importance, medium
importance, low importance, and very low importance (see e.g. Fu et al., 2010), using the natural breaks data classification of ArcMap.

proposed network (Fig. 6) summed 240 patches (153.24 km2 ) in forestry patches (40.45 km2 and a mean area of 1.61 km2 ) mixed
total. Substitution actions were proposed for 28 patches (40.96 km2 with a variety of natural species such as Quercus robur, Quercus
and a mean area of 1.46 km2 ) of Eucalyptus globulus and Pinus pyrenaica, Quercus ilex, Castanea sativa, and Corylus avellana among
radiata spp. Restoration actions were proposed for 25 plantation others, probably all relict remains of the native forest.
512 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515

Fig. 5. Results of dIIC and dPC at various dispersal distances in the 2a and 2b scenarios.

Table 7 Discussion
dIIC and dPC values of 2a and 2b scenarios.

2000 5000 10,000 15,000 25,000 Plantation forestry contributes to connectivity improvement
dIIC in natural forest networks, and increases native habitat availabil-
Max 25.08 58.98 36.14 30.60 28.42 ity for forest dwelling species when the replanted species are
Scenario 2a (with
highways)
Mean 0.710 1.109 1.122 1.115 1.133 autochthonous (Crouzeilles et al. 2014). We analysed landscape
Sum 338.22 528.35 534.39 531.20 539.49 connectivity considering different scenarios involving both patches
Scenario 2b Max 25.06 62.48 36.08 31.55 29.25
of natural forest and re-planted ones. We found clear differences
(without Mean 0.710 1.240 1.330 1.148 1.156
highways) Sum 338.32 590.55 633.17 546.56 550.44 among patches depending on their size, location and the dispersal
distance under study. Selection of the most suitable patches in
dPC
Max 33.02 24.17 27.04 27.14 26.34
terms of connectivity in forest replantation schemes may be crucial
Scenario 2a (with when attempting specific conservation actions and is a paramount
Mean 0.816 0.959 1.086 1.137 1.170
highways)
Sum 388.62 456.93 517.34 541.27 556.98 attribute of forest management (García-Feced et al. 2011; Pascual-
Scenario 2b Max 32.88 25.23 29.85 29.63 28.20 Hortal & Saura, 2008; Saura et al. 2011b).
(without Mean 0.831 1.037 1.213 1.248 1.240
highways) Sum 395.64 493.67 577.51 594.18 590.69

Fig. 6. Proposed forest landscape network.

 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515 513

Variation in plantation patch importance for forest connectivity concentration of routes crossing the viaducts promoted the fun-
damental role of various patches located close to them. This is
Higher habitat availabilities and good connectivity between an important finding and may have major implications when
patches were beneficial for most species (Laita et al. 2010; Rubio & considering future fauna crossing constructions, since it suggests
Saura, 2012). In our study, large patches and groups of patches with that those patches close to fauna crossings should be protected
varying sizes but with concentrated spatial distributions were the and conserved with special attention. Similarly, it suggests that
most relevant in terms of connectivity for medium-sized and large fauna crossings devoid of vegetation should be modified by re-
mammals, as shown by other authors (Liu et al. 2014a; Shathala planting forest or bush species in the immediate vicinity. This
Devi et al. 2013). We considered various dispersal distances in will attract forest-dwelling species and favour their movements
order to include the broadest part of the existing dispersal needs through appropriate structures (Kaphegyi et al. 2013). The influ-
(Baguette et al. 2013; García-Feced et al. 2011; Gurrutxaga et al. ence of viaducts in our results also indicates that when analysing
2011; Saura et al. 2011b). We found that centrally located patches connectivity, it is crucial to conduct a precise delimitation of the
were more relevant in terms of connectivity maintenance within key elements of the matrix (such as highways), if a correct analysis
the forest network for longer dispersal distances, since they acted of animal movements is to be achieved.
as stepping-stone in the majority of links established between The connectivity interruption imposed by the highway in our
large peripheral patches. Thus, our results indicate that despite study area was mainly due to the absence of specific faunal
differences in patch relevance, the whole natural network should crossings, a common feature in old highways, which were built
be preserved, given its low prevalence in the study area. On the without considering ecological permeability issues (Gurrutxaga
other hand, plantation patches minimise distances between the et al. 2011). Sometimes, highway permeability has been overes-
high natural value areas located in the north, whilst thickening the timated (Kaphegyi et al. 2013), since fauna may use non-specific
forest network at the south of the study area. In this way, the pro- crossings, and often trespass the highway, causing faunal run over
posed (natural plus replanted) network may provide a more robust and road safety problems. Therefore, it would be appropriate to
network of patches and a higher resilience in case of future habi- improve aspects such as perimeter fencing of these infrastructures,
tat losses or local extinctions (Laita et al. 2011; Piquer-Rodríguez or propose other measures aimed at hampering animal access to the
et al. 2012; Saura et al. 2011b), mainly because connectivity is highways, as has been recommended by various authors (Kaphegyi
better shared (i.e. less concentrated) among a larger number of et al. 2013 and references therein). In this way, animals would be
patches. forced to travel through specific (or non-specific) faunal crossings
Regardless of connectivity improvements with plantation patch (Mata et al. 2005; Rodríguez et al. 1996; Saura et al. 2011a).
addition, relevant patches and their spatial location were simi- By using a combination of several connectivity metrics based
lar at all dispersal distances. García-Feced et al. (2011) also found on graph-theory (Shathala Devi et al. 2013; Szabó et al. 2012), we
that key patches for connectivity were those which had previously showed how the application of these tools may lead to a substantial
been identified as critical when prioritizing patches for reforesta- leap forwards in identifying key patches in terms of forest network
tion purposes. This highlights the importance of specific patches improvement, as well as those elements of the landscape matrix
for good and bad dispersers and that these specific nodes should be which are most determinant for animal dispersal. This is impor-
considered no matter the management measures under discussion. tant when establishing investment priorities and protection and
Based on our results and considering the particularities of the study landscape improvement measures, since resources are currently
area, a good management strategy would be to replant the area situ- limited. By improving connectivity of the forest network species
ated north of the highway with autochthonous species (very scarce conservation is strengthened, which is a priority in environmental
nowadays), favouring larger and well-connected populations also planning.
towards the south of the area, which is mountainous and richer in
natural forest patches.
Acknowledgements
Highways hamper species dispersal but may promote the role of
The authors would like to acknowledge the revision of the
specific patches
English manuscript to N.L. Arroyo Hailuoto. We also thank J. Sáinz-
Martínez for his support with the elaboration of the cartographic
The highway crossing the study area was a crucial factor when
material.
analysing landscape connectivity. As shown in the two pairs of
scenarios considered (with and without highway), those scenarios
including the highway allowed a lower number of connections (e.g. References
Szabó et al. 2012) and a general increase in the distance between
pairs of patches, which obtained lower values in terms of connectiv- Baguette, M., Blanchet, S., Legrand, D., Stevens, V. M., & Turlure, C. (2013). Individual
dispersal, landscape connectivity and ecological networks. Biological Reviews,
ity importance. This suggests that the highway severely hampered 88, 310–326.
species dispersal, which is in accordance with other studies (Fu Blanco, J. C., Cortés, Y., & Virgós, E. (2005). Wolf response to two kinds of bar-
et al. 2010; Liu et al. 2014c). Moreover, in so far as the most rele- riers in an agricultural habitat in Spain. Canadian Journal of Zoology, 83,
312–323.
vant highway effects were found at the longest distances for both Brockerhoff, E. G., Berndt, L. A., & Jactel, H. (2005). Role of exotic pine forest in
indices, it can be deduced that those species more acutely affected the conservation of the critically endangered ground beetle Holcaspis brevicula
by the highway are those with better dispersal abilities, that is, (Coleoptera: Carabidae). New Zealand Journal of Ecology, 29, 37–43.
Brockerhoff, E. G., Ecroyd, C. E., Leckie, A. C., & Kimberley, M. O. (2003). Diversity
those able to cover larger distances, and thus more susceptible of
and succession of adventive and indigenous vascular understory plants in Pinus
crossing the highway in their movements. radiata plantation forest in New Zealand. Forest Ecology and Management, 185,
Notwithstanding the fact that highways create larger dispersal 307–326.
Brockerhoff, E. G., Jactel, H., Parrotta, J. A., Quine, C. P., & Sayer, J. (2008). Plantation
needs for most species, we observed that some patches became
forests and biodiversity: Oxymoron or opportunity? Biodiversity Conservation,
more important in terms of connectivity in those scenarios includ- 17, 925–951.
ing the highway. This was probably due to the three viaducts Calviño-Cancela, M., Rubido-Bará, M., & van Etten, E. J. B. (2012). Do eucalypt
acting as bottlenecks in the otherwise hermetic highway. These plantations provide habitat for native forest biodiversity? Forest Ecology and
Management, 270, 153–162.
free resistance stretches were useful as catalysers of the majority Carnus, J.-M., Parrotta, J., Brockerhoff, E. G., Arbez, M., Jactel, H., Kremer, A., et al.
of least cost paths (as identified by Cuhsman et al. 2013), while the (2006). Planted forest and biodiversity. Journal of Forestry, 104(2), 65–77.
514 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515

Crouzeilles, R., Prevedello, J. A., Lima-Figueiredo, M. S., Lorini, M. L., & Viveiros-Grelle, Li, T., Shilling, F., Thome, J., Li, F., Schott, H., Boynton, R., et al. (2010). Fragmentation
C. E. (2014). The effects of the number, size and isolation of patches along a of China’s landscape by roads and urban areas. Landscape Ecology, 25, 839–853.
gradient of native vegetation cover: How can we increment habitat availability? Liu, S., Deng, L., Chen, L., Li, J., Dong, S., & Zhao, H. (2014). Landscape network
Landscape Ecology, 29, 479–489. approach to assess ecological impacts of road projects on biological conserva-
Cuhsman, S. A., Lewis, J. S., & Landguth, E. L. (2013). Evaluating the intersection of a tion. Chinese Geographical Science, 24(1), 5–14.
regional wildlife connectivity network with highways. Movement Ecology, 1(12), Liu, S., Deng, L., Dong, S., Zhao, Q., Yang, J., & Wang, C. (2014). Landscape connectiv-
1–11. ity dynamics based on network analysis in the Xishuangbanna Nature Reserve,
Dean, W. R. J., & Milton, S. J. (2003). The importance of roads and road verges for rap- China. Acta Oecologica, 55, 66–77.
tors and crows in the Succulent and Nama-Karoo, South Africa. Ostrich, 74(3–4), Liu, S., Deng, L., Zhao, Q., DeGloria, S. D., & Dong, S. (2011). Effects of road network
181–186. on vegetation pattern in Xishuangbanna, Yunnan Province, Southwest China.
Decout, S., Manel, S., Miaud, C., & Luque, S. (2012). Integrative approach for Transportation Research Part D, 16, 591–594.
landscape-based graph connectivity analysis: A case study with the common Liu, S., Dong, Y., Deng, L., Liu, Q., Zhao, H., & Dong, S. (2014). Forest fragmentation
frog (Rana temporaria) in human-dominated landscapes. Landscape Ecology, 27, and landscape connectivity change associated with road network extension and
267–279. city expansion: A case study in the Lancang River Valley. Ecological Indicators,
Di Giulio, M., Holderegger, R., & Tobias, S. (2009). Effects of habitat and landscape 36, 160–168.
fragmentation on human and biodiversity in densely populated landscapes. Liu, S. L., Cui, B. S., Dong, S. K., Yang, Z. F., Yang, M., & Holt, K. (2008). Evaluating
Journal of Environmental Management, 90, 2959–2968. the influence of road networks on landscape and regional ecological risk – A
Dodd, N. L., Gagnon, J. W., Boe, S., & Schweinsburg, R. E. (2007). Assessment of elk case study in Lancang River Valley of Southwest China. Ecological Engineering,
highway permeability by using global positioning system telemetry. Journal of 34, 91–99.
Wildlife Management, 71(4), 1107–1117. Mata, C., Hervás, I., Herranz, J., Suárez, F., & Malo, J. E. (2005). Complementary use by
Fabiao, A., Martins, M. C., Cerveira, C., Santos, C., Lousa, M., Madeira, M., et al. (2002). vertebrates of crossing structures along a fenced Spanish motorway. Biological
Influence of soil and organic residue management on biomass and biodiversity Conservation, 214, 397–405.
of understory vegetation in a Eucalyptus globulus Labill plantation. Forest Ecology McRae, B. H., & Kavanagh, D. M. (2011). Linkage mapper connectivity analy-
and Management, 171, 87–100. sis software. Seattle, WA: The Nature Conservancy. Available at http://www.
Fahrig, L. (2007). Non-optimal animal movement in human-altered landscapes. circuitscape.org/linkagemapper
Functional Ecology, 21, 1003–1015. MMA. (2002). Mapa Forestal de España 1:50.000. Madrid, España: Ministerio de Medio
Fearnside, P. M. (2005). Deforestation in Brazilian Amazonia: History, rates, and Ambiente, Organismo Autónomo de Parques Nacionales.
consequences. Conservation Biology, 19(3), 680–688. Pascual-Hortal, L., & Saura, S. (2006). Integrating landscape connectivity in broad-
Fischer, J., & Lindenmayer, D. B. (2007). Landscape modification and habitat frag- scale forest planning: A methodology based on graph structures and habitat
mentation: A synthesis. Global Ecology and Biogeography, 16, 265–280. availability indices. In R. Lafortezza, & G. Sanesi (Eds.), Patterns and processes in
Forman, R. T. T., & Alexander, L. E. (1998). Roads and their major ecological effects. forest landscapes. Consequences of human management (pp. 111–116).
The Annual Review of Ecology, Evolution, and Systematics, 29, 207–231. Pascual-Hortal, L., & Saura, S. (2008). Integrating connectivity in broad-scale forest
Forman, R. T. T., & Deblinger, R. D. (2000). The ecological road-effect zone planning through a new graph-based habitat availability methodology: Appli-
of a Massachusetts (U.S.A.) suburban highway. Conservation Biology, 14(1), cation to capercaillie (Tetrao urogallus) in Catalonia (NE Spain). European Journal
36–46. of Forest Research, 127, 23–31.
Forman, R. T. T., Reineking, B., & Hersperger, A. M. (2002). Road traffic and nearby Pereira, M., Segurado, P., & Neves, N. (2011). Using spatial network structure in
grassland bird patterns in a suburbanizing landscape. Environmental Manage- landscape management and planning: A case study with pond turtles. Landscape
ment, 29(6), 782–800. and Urban Planning, 100, 67–76.
Fu, W., Liu, S., Degloria, S. D., Dong, S., & Beazley, R. (2010). Characterizing the Piquer-Rodríguez, M., Kuemmerle, T., Alcaraz-Segura, D., Zurita-Milla, R., & Cabello,
fragmentation-barrier effect of road networks on landscape connectivity: A case J. (2012). Future land use effects on the connectivity of protected area networks
study in Xishuangbanna, Southwest China. Landscape and Urban Planning, 95, in southeastern Spain. Journal for Nature Conservation, 20(6), 326–336.
122–129. Poore, M. E. D., & Fries, C. (1985). The ecological effects of eucalyptus. Food and Agri-
Galpern, P., Manseau, M., & Fall, A. (2011). Patch-based graphs of landscape con- culture Organization of the United Nations forestry paper 59.
nectivity: A guide to construction, analysis and application for conservation. Rodríguez, A., Crema, G., & Delibes, M. (1996). Use of non-wildlife passages across
Biological Conservation, 144, 44–55. a high speed railway by terrestrial vertebrates. Journal of Applied Ecology, 33,
García-Feced, C., Saura, S., & Elena-Rosselló, R. (2011). Improving landscape connec- 1527–1540.
tivity in forest districts: A two-stage process for prioritizing agricultural patches Rubio, L., & Saura, S. (2012). Assessing the importance of individual habitat patches as
for reforestation. Forest Ecology and Management, 261, 154–161. irreplaceable connecting elements: An analysis of simulated and real landscape
García-Martino, F. (1862). Bosquejo dasográfico de la provincia de Santander 1:250.000. data. Ecological Complexity, 11, 28–37.
Madrid, España: Junta General de Estadística. Rubio, L., Rodríguez-Freire, M., Mateo-Sánchez, M. C., Estreguil, C., & Saura, S. (2012).
Geneletti, D. (2006). Some common shortcomings in the treatment of impacts of lin- Sustaining forest landscape connectivity under different land cover change sce-
ear infrastructures on natural habitats. Environmental Impact Assessment Review, narios. Forest Systems, 2(21), 223–235.
26, 257–267. Rytwinski, T., & Fahrig, L. (2013). Why are some animal populations unaffected or
Gurrutxaga, M., Rubio, L., & Saura, S. (2011). Key connectors in protected forest positively affected by roads? Oecologia, 173, 1143–1156.
area networks and the impact of highways: A transnational case study from Saura, S., & Pascual-Hortal, L. (2007). A new habitat availability index to integrate
the Cantabrian Range to the Western Alps (SW Europe). Landscape and Urban connectivity in landscape conservation planning: Comparison with existing
Planning, 101, 310–320. indices and application to a case study. Landscape and Urban Planning, 83,
Harper, G. J., Steininger, M. K., Tucker, C. J., Juhn, D., & Hawkins, F. (2007). Fifty 91–103.
years of deforestation and forest fragmentation in Madagascar. Environmental Saura, S., & Torné, J. (2009). Conefor Sensinode 2.2: A software package for quantify-
Conservation, 34(4), 325–333. ing the importance of habitat patches for landscape connectivity. Environmental
Humphrey, J. W. (2005). Benefits to biodiversity from developing old-growth con- Modelling & Software, 24, 135–139.
ditions in British upland spruce plantations: A review and recommendation. Saura, S., Vogt, P., Velázquez, J., Hernando, A., & Tejera, R. (2011). Key structural
Forestry, 78(1), 33–53. forest connectors can be identified by combining landscape spatial pattern and
Humphrey, J. W., Newton, A. C., Peace, A. J., & Holden, E. (2000). The importance of network analyses. Forest Ecology and Management, 262, 150–160.
conifer plantations in northern Britain as a habitat for native fungi. Biological Saura, S., Estreguil, C., Mouton, C., & Rodríguez-Freire, M. (2011). Network anal-
Conservation, 96, 241–252. ysis to assess landscape connectivity trends: Application to European Forests
IGN. (2006). Instituto Geográfico Nacional Cartography BCN200. (1990–2000). Ecological Indicators, 11, 407–416.
Kaphegyi, T. A. M., Dees, M., Zlatanova, D., Ueffing, C., Dutsov, A., & Kaphegyi, U. Shathala Devi, B. S., Murthy, M. S. R., Debnath, B., & Jha, C. S. (2013). Forests patch con-
(2013). Rapid assessment of linear transport infrastructure in relation to the nectivity diagnostics and prioritization using graph theory. Ecological Modelling,
impact on landscape continuity for large ranging mammals. Biodiversity Conser- 251, 279–287.
vation, 22, 153–168. Skole, D., & Tucker, C. (1993). Tropical deforestation and habitat fragmentation in
Kramer-Schadt, S., Revilla, E., Wiegand, T., & Breitenmoser, U. (2004). Fragmented the Amazon: Satellite data from 1978 to 1988. Science, 260(5116), 1905–1910.
landscapes, road mortality and patch connectivity: Modeling influences on the Spellerberg, I. F. (1998). Ecological effects of roads and traffic: A literature review.
dispersal of Eurasian lynx. Journal of Applied Ecology, 41, 711–723. Global Ecology and Biogeography Letters, 7(5), 317–333.
Laita, A., Mönkkönen, M., & Kotiaho, J. S. (2010). Woodland key habitat eval- Szabó, S., Novak, T., & Elek, Z. (2012). Distance models in ecological network man-
uated as part of a functional reserve network. Biological Conservation, 143, agement: A case study of path connectivity in a grassland network. Journal for
1212–1227. Nature Conservation, 20, 293–300.
Laita, A., Kotiaho, J. S., & Mönkkönen, M. (2011). Graph-theoretic connectivity Taylor, P. D., Fahrig, L., Henein, K., & Merriam, G. (1993). Connectivity is a vital
measures: What do they tell us about connectivity? Landscape Ecology, 26, element of landscape structure. Oikos, 68, 571–573.
951–967. Taylor, P. D., Fahrig, L., & With, K. A. (2006). Landscape connectivity: A return to the
Lamb, D., Erskine, P. D., & Parrotta, J. A. (2005). Restoration of degraded tropical forest basics. Conservation Biology, 14, 29–43.
landscapes. Science, 310(5754), 1628–1632. Tischendorf, L., & Fahrig, L. (2000). How should we measure landscape connectivity?
Langevelde, F., Dooremalen, C., & Jaarsma, C. F. (2009). Short communication: Traffic Landscape Ecology, 15, 633–641.
mortality and the role of minor roads. Journal of Environment Management, 90, Urban, D., & Keitt, T. (2001). Landscape connectivity: A graph-theoretic perspective.
660–667. Ecology, 82(5), 1205–1218.
 S. Nogués, A. Cabarga-Varona / Journal for Nature Conservation 22 (2014) 504–515 515

Waller, J. S., & Servheen, C. (2005). Effects of transportation infrastructure on Yu, D., Xun, B., Shi, P., Shao, H., & Liu, Y. (2012). Ecological restoration planning based
grizzly bears in northwestern Montana. Journal of Wildlife Management, 69(3), on connectivity in an urban area. Ecological Engineering, 46, 24–33.
985–1000. Ziólkowska, E., Ostapowicz, K., Kuemmerle, T., Perzanowski, K., Radeloff, V. C., &
Watts, K., & Handley, P. (2010). Developing a functional connectivity indica- Kozak, J. (2012). Potential habitat connectivity of European bison (Bison bonasus)
tor to detect change in fragmented landscapes. Biological Indicators, 10, in the Carpathians. Biological Conservation, 146, 188–196.
552–557.