Content uploaded by Rodney van der Ree
Author content
All content in this area was uploaded by Rodney van der Ree on Mar 16, 2018
Content may be subject to copyright.
Permeability of the urban matrix to arboreal gliding
mammals: Sugar gliders in Melbourne, Australia
FIONA M. CARYL,1KATRINA THOMSON1,2 AND RODNEY VAN DER REE1,2*
1School of Botany, University of Melbourne, Melbourne, Victoria 3010, Australia
(Email: rvdr@unimelb.edu.au), and 2Australian Research Centre for Urban Ecology, Royal Botanic
Gardens Melbourne, C/- School of Botany, University of Melbourne, Melbourne, Victoria, Australia
Abstract Habitat corridors that facilitate functional connectivity are a fundamental component of wildlife
conservation in fragmented landscapes. However, the landscape matrix separating suitable habitat is not uniformly
impermeable to movement and management to increase matrix permeability could be an alternative means to
maintain connectivity. Gliding mammals are particularly sensitive to fragmentation because their movements are
constrained by glide distance thresholds. Populations of gliders in cities are at risk of being isolated by increasing
habitat loss and urban development, yet little is known about how the urban matrix affects glider movement. Here
we investigate how the level of urbanization and tree cover in the matrix influence matrix permeability to sugar
gliders (Petaurus breviceps) within suburban forest reserves.Twenty-two sugar gliders were radio-tracked over winter
and summer at four reserves. Boundary crossing behaviour was measured as the number of times each glider
crossed into the matrix, and matrix permeability was determined as the maximum distance travelled by gliders into
the matrix. The majority of gliders (81%) were located in the matrix at least once, and rates of boundary crossing
were consistent across urbanization and tree cover levels. Matrix permeability was negatively affected by matrix
urbanization, but not by matrix tree cover, and no interaction effects were found. Although distances travelled by
gliders into the matrix did not exceed 180 m, they were comparable with typical movement distances by gliders in
reserves. Our results demonstrate that the urban matrix can provide suitable habitat for gliding mammals to move
and forage, but that increased urbanization may inhibit glider use of the matrix irrespective of tree cover. This
finding has implications for conservation planning and suggests that structurally connected areas may not be used
if movement behaviour is inhibited. Conversely, management of matrix permeability could be used to maintain
connectivity without needing to construct physical corridors.
Key words: animal movement, functional connectivity, matrix management, urban landscape, wildlife corridor.
INTRODUCTION
The quality, extent and spatial configuration of habitat
patches are critical influences on the persistence of
native fauna within fragmented landscapes (Forman
1995; Fahrig 2003).Yet because of its effect on inter-
patch movement (i.e. functional connectivity), the
matrix in which habitat patches are embedded can also
influence the abundance and distribution of fauna
within landscapes (Prevedello & Vieira 2010; Watling
et al. 2011). Functional connectivity is species-specific
and relates movement behaviours to patterns of
landscape structure (Tischendorf & Fahrig 2000;
Calabrese & Fagan 2004; Taylor et al. 2006). The per-
meability of matrix habitats to animal movement is
therefore often influenced by their structural similarity
with preferred habitats (Prevedello & Vieira 2010).
Where structural contrast with preferred habitats
is high, the matrix acts as a barrier to movement;
however, a low contrast matrix may be sufficiently
permeable to allow dispersal among habitats and
populations (Ricketts 2001).
Management aimed at increasing matrix permeabil-
ity for wildlife could be a valuable tool for restoring
landscape connectivity (Taylor et al. 2006; Noss 2007;
Schmiegelow 2007). However, understanding matrix
permeability requires detailed information about
species-specific habitat selection, dispersal ability and
movement, all of which require extensive, and often
labour-intensive, fieldwork to attain (Calabrese &
Fagan 2004; Baguette & van Dyck 2007). Because
structural connectivity (which describes the physical
relationships between habitat patches such as inter-
patch distance) is relatively simple to quantify, connec-
tivity research has focused on linking patches with
corridors, while the influence of the matrix has
largely been ignored (Beier & Noss 1998; Noss 2007;
*Corresponding author.
Accepted for publication October 2012.
Austral Ecology (2013) 38, 609–616
bs_bs_banner
© 2012 The Authors doi:10.1111/aec.12006
Austral Ecology © 2012 Ecological Society of Australia
Schmiegelow 2007; Taylor et al. 2006).There are con-
siderable economic costs and trade-offs associated
with establishing physical corridors along with an
ongoing debate as to their efficacy (e.g. Simberloff
& Cox 1987; Simberloff et al. 1992; Saunders 2007).
Management of the matrix could offer an alterna-
tive (or augmentative) means to promote movement
through landscapes; however, determining its effec-
tiveness first requires greater understanding of what
influences matrix permeability (Taylor et al. 2006;
Noss 2007; Schmiegelow 2007).
Gliding mammals are particularly sensitive to frag-
mentation because their ability to move through the
landscape is constrained by the extent of tree cover (van
der Ree et al. 2003; Goldingay & Taylor 2009). Gaps
in tree cover that exceed glide distance thresholds are
effectively barriers to movement for gliding species that
rarely venture onto the ground (van der Ree et al. 2003;
Ball & Goldingay 2008). More than 60 glider species
are threatened worldwide (Goldingay 2000), meaning
there is considerable interest in establishing means
to maintain landscape connectivity for this group
(e.g. Selonen & Hanski 2003; Ball & Goldingay 2008;
Ritchie et al. 2009). In Australia, where four out of six
species of glider are threatened in some part of their
range (Petaurus australis; Petaurus norfolcensis; Petaurus
gracilis; Petauroides volans; Goldingay 2000), much
research into the effects of fragmentation has con-
centrated on gliders in agricultural systems (Suckling
1984; Lindenmayer et al. 1999; van der Ree 2002; van
der Ree & Bennett 2003; van der Ree et al. 2003). But
as urbanization becomes an increasing threat to species
survival there is growing interest in glider ecology
within urban areas (Goldingay & Sharpe 2004; Taylor
& Goldingay 2009; Brearley et al. 2010, 2011a,b).
Despite this, our understanding of glider dispersal and
what makes the urban matrix permeable to gliders
remains poor (Goldingay et al. 2006).
How an animal responds to an edge provides a good
indication of what influences their landscape perme-
ability (Stamps et al. 1987; Haddad 1999). Squirrel
gliders (Petaurus norfolcensis) in urban areas demon-
strate a strong response to high contrast (hard) edges
between preferred habitat and matrix (as evidenced
by lower abundances and smaller home ranges) but
are somewhat tolerant of low contrast (soft) edges
(Brearley et al. 2010, 2011b). Boundary crossing and
the proclivity of an animal to move through the matrix
are precursory behaviours that lead to dispersal (Cala-
brese & Fagan 2004; Baguette & van Dyck 2007).
Knowledge of what influences these behaviours will
provide better understanding of what makes habitat
edges ‘soft’ and thus what drives matrix permeability
for gliding mammals. In this paper we determine how
movement of sugar gliders (Petaurus breviceps) into the
urban matrix is influenced by the level of tree cover
and urbanization in the matrix.
METHODS
Study species
The sugar glider (Petaurus breviceps) is a small (104–119 g),
nocturnal, gliding mammal that is widespread in the wood-
lands of eastern and northern Australia (Quin 1995; Gold-
ingay & Jackson 2004). While not currently threatened, the
species is similar to other threatened species of glider in that
they are limited by their dispersal abilities and rely on mature
eucalypt communities to fulfil their specific diet and nesting
requirements (Suckling 1984; Quin 1995; Goldingay &
Jackson 2004).
Study area and sites
The study was conducted in the south-east suburbs of Mel-
bourne, approximately 16–23 km from the central business
district (CBD: Fig. 1). Melbourne is the second largest city
in Australia with a human population of approximately 4.1
million spread over an area 8806 km2(Australian Bureau of
Statistics 2012). Prior to European settlement in 1835, the
area supported a wide variety of plant communities domi-
nated by dry sclerophyll forests (Oates & Taranto 2001).
These have since been cleared and modified such that only
1.7% of the vegetation in the inner suburbs (those within 10
km of the central business district) is from native remnants,
which increases to approximately 26% of vegetation in the
outer suburbs (van der Ree, 2005, unpubl. data). Our study
area in the eastern suburbs supports approximately 5.7%
natural vegetation occurring as a few large (>100 ha) patches
and numerous small patches (<20 ha: van der Ree, 2005,
unpubl. data).Tree canopy layers in the area were dominated
by Eucalyptus species (e.g. Eucalyptus macrorhyncha, Eucalyp-
tus melliodora, Eucalyptus obliqua, Eucalyptus viminalis), while
subcanopy vegetation primarily consisted of Acacia species
(e.g. Acacia dealbata, Acacia mearnsii, Acacia melanoxylon).
The understorey had a variable mixture of native and intro-
duced grasses, forbs and shrubs.
We selected four forest reserves (the 100 Acres, Jells Park,
Valley Reserve and Damper Creek) occupied by gliders that
occurred closest to the centre of Melbourne (Table 1). All
reserves were surrounded by urban land uses (predominantly
low- to medium-density residential housing), and were adja-
cent to watercourses. Average tree cover within reserves was
52 ⫾8% (range 34–72%). The landscape context of each
reserve was used to group gliders by level of matrix urbani-
zation and tree cover (see Spatial data below).
Animal capture
Animals were trapped at the 100 Acres Reserve, but were
obtained from nest boxes that had been installed specifically
for sugar gliders at Jells Park, Valley Reserve and Damper
Creek. Trapping at 100 Acres was conducted by bracketing
20 small mammal traps (Elliott size A, 9 ¥9¥33 cm) to tree
trunks at a height of approximately 4 m, located randomly at
50 m intervals throughout the reserve. Traps were protected
from adverse weather with plastic and left in situ for 10 nights
610 F. M. CARYL ET AL.
© 2012 The Authorsdoi:10.1111/aec.12006
Austral Ecology © 2012 Ecological Society of Australia
while baited with a mixture of rolled oats, peanut butter and
honey. A trail of diluted honey was sprayed up tree trunks as
an attractant.Traps were checked at dawn each morning and
any captured animals removed for processing. Trapped
gliders, and those removed from nest boxes, were weighed,
sexed, aged (following Suckling 1984), and had their repro-
ductive condition assessed. Animals were then radio-collared
and released at the point of capture. Collared gliders were
fitted with a tuned-loop single-stage transmitter (Sirtrack,
New Zealand) operating within frequencies of 150–
151 MHz. Each transmitter and collar weighed between 4.0
and 5.5 g, equating to <5% of an adult sugar glider’s body
weight. Collars were covered in heat shrink and a strip of
reflective tape glued to each transmitter to assist location in a
spotlight beam at night.
Radio-tracking
Radio-tracking was conducted on foot in summer (2 Novem-
ber 2005 to 25 February 2006) and winter (7 June 2006 and
15 October 2006) using a collapsible three-element Yagi
antenna and a Regal 2000 or Australis 26 k receiver (Titley
Electronics, Ballina). The order in which parks were visited,
and animals tracked, was selected randomly and nights near
the full moon were avoided. One location fix was collected
Fig. 1. Location of four forest reserves used in the study within Australia (top left) and Melbourne (bottom left). Forest
remnants (white) shown against urban matrix (grey) and roads (lines: right). AC, the 100 acres; DC, Damper Creek; JP, Jells
Park; VR, Valley Reserve.
Ta b l e 1 . Study site information. The number of sugar gliders tracked at each reserve is provided along with the distance of the
centre of each reserve to Melbourne’s central business district (CBD). Reserve area, perimeter–area (PA) ratio and extent of tree
cover within reserve boundaries are provided. Building, human and road density within a 260-m buffer from the reserve were
highly correlated (r=0.92–0.99) so were used to group reserves as having high (H) or low (L) matrix urbanization (URBAN),
while percentage tree cover within buffers grouped reserves with high or low tree cover (TREE)
Reserve name 100 Acres Damper Creek Jells Park Valley Reserve
Gliders (n)6 6 4 6
Distance from CBD (km) 17 23 16 23
Area (ha) 40 15 130 16
PA ratio (m ha-1) 64 287 57 160
Within-reserve tree cover (%) 72 53 34 50
Building density (nha-1) 2.8 12.4 4.7 12
Human density (nha-1) 4 23 6 20
Road density (m ha-1) 50 120 29 100
URBAN (Group) L H L H
Matrix tree cover (%) 49 23 17 37
TREE (Group) H L L H
PERMEABILITY OF URBAN MATRIX TO GLIDERS 611
© 2012 The Authors doi:10.1111/aec.12006
Austral Ecology © 2012 Ecological Society of Australia
per animal per night between sunset and 3 a.m. while animals
were active. This ensured statistical independence of fixes
and allowed all gliders to be located once each night. Up to
10 min was spent searching for each animal using a 12 V
50 W spotlight with red filter. All fixes were recorded with a
hand-held GPS unit (⫾5 m). When gliders visited residential
yards, triangulation from two or three sides of the yard deter-
mined the animal’s location. Residential yards typically con-
tained few large trees so identifying which tree the animal
was in was relatively simple.
Spatial data
Bushland reserve boundaries were delineated using the Mel-
bourne Open Space Map (Leary & McDonnell 2001) within
ArcViewGIS (ESRI Inc., California, USA). Glider locations
were defined as being within the matrix when they occurred
>5 m outside reserve boundaries (to account for our mean
GPS error). Reserve boundaries were buffered by a distance
of 260 m (the mean maximum distance found between any
pair of locations for each glider) in which properties of the
matrix were then measured. Patterns of urban landscape
form and structure can vary depending on the measure used
to quantify urbanization (Hahs & McDonnell 2006). To
reflect the extent of urbanization surrounding each reserve
we therefore measured the density of roads, buildings, and
humans within reserve buffers using data from Land Victoria
(2001) and the Australian Bureau of Statistics (2003). As
these three variables were highly correlated (r=0.92–0.99),
gliders within reserves could be classified according to their
‘matrix urbanization’ as high (>100 m roads ha-1,>10 build-
ings ha-1,>20 people ha-1:n=12) or low (<50 m roads ha-1,
<5 buildings ha-1,<10 people ha-1:n=10).
The level of tree canopy cover in reserve buffers was deter-
mined from per cent tree cover usingTree25 (Department of
Sustainability and Environment 2007). Tree25 is a mapped
presence/absence layer derived from a combination of digital
classification and visual interpretation of SPOT Panchro-
matic imagery with a 10 m pixel size. In this layer, tree cover
is defined as woody vegetation greater than 2 m in height and
with a crown cover (foliar density) greater than 10%. Tree
cover within the surrounding matrix was used to classify
gliders within reserves by their ‘matrix tree cover’ as high
(>35% tree cover, n=12) or low (<25%, n=10). These
thresholds meant that matrix with high tree cover had a level
of tree cover that was equal to or higher than that within the
reserve with the lowest (internal) tree cover. Similarly, matrix
with low tree cover had up to 50% less tree cover than that
within the reserve with the lowest (internal) tree cover.
Resultant cross-tabulated sample sizes (number of gliders)
for high (H) and low (L) matrix urbanization (U) and matrix
tree cover (T) were HUHT =6; HULT =6; LUHT =6;
LULT =4.
The distance from the centre of each glider’s core home
range to the nearest reserve boundary was used to determine
if matrix permeability was related to proximity of home
ranges to reserve edges. We defined core home range using
the Home Range Extension for ArcView (Rodgers & Carr
1998) to estimate the minimum convex polygon (MCP:
Harris et al. 1990) formed by 50% of locations for each glider
(White & Garrott 1990). As core ranges reflected where
animals spent 50% of their activity, we used the mean dis-
tance between core locations of each glider to determine if
movements into the matrix were comparable with typical
movement distances.
Statistical analysis
All statistical analyses were performed using SPSS Statistics
for Windows, Version 17.0 (SPSS Inc., Chicago, IL, USA)
and all averages are presented as untransformed means ⫾1
standard error to aid interpretation. For each glider we cal-
culated two responses: (i) boundary crossing frequency (the
number of times each glider crossed into the urban matrix
divided by the total number of active locations for that
animal); and (ii) matrix permeability (the maximum distance
travelled by each glider into the matrix). Boundary crossing
frequency was transformed prior to analysis to improve nor-
mality as 1/(y+1), while matrix permeability was trans-
formed as ln(y). Glider ranging behaviour will vary between
sexes and among seasons (Sharpe & Goldingay 2007);
however, we pooled our data because our sample size was too
small to analyse the influence of these variables.
Gliders were grouped according to the level of urbaniza-
tion and tree cover in the matrix surrounding their reserves.
A general linear model (two-way factorial anova) was then
used to test for single and interaction effects of matrix
urbanization and tree cover on glider boundary crossing fre-
quency and matrix permeability. Homogeneity of residual
variances among groups was checked using Levene’s test. A
simple linear regression was used to determine if matrix
permeability was influenced by proximity of core ranges to
reserve boundaries.
RESULTS
A total of 535 locations were obtained for 22 (18
female, 4 male) sugar gliders, with approximately
equal numbers of locations per animal (24.3 ⫾6.9),
98 of which occurred within the matrix. The majority
of gliders (81%) were located in the matrix at least
once during the study, and were often observed forag-
ing there. In summer, four gliders (three females and
one male) were not recorded in the matrix and were
therefore removed from analyses of matrix permeabil-
ity; however, these individuals were retained for analy-
ses of boundary crossing.
We found no interactive effects of matrix urbani-
zation and tree cover on glider boundary crossing
behaviour or matrix permeability (Table 2). Although
a higher proportion of glider locations tended to
occur more frequently in matrix with high tree cover
(23% of locations) than low tree cover (11% of loca-
tions: Fig. 2a), and in matrix with low urbanization
(23% of locations) than high urbanization (13% of
locations: Fig. 2b), these effects were not statistically
significant (Table 2). Matrix permeability was not
influenced by matrix tree cover (mean difference
612 F. M. CARYL ET AL.
© 2012 The Authorsdoi:10.1111/aec.12006
Austral Ecology © 2012 Ecological Society of Australia
6.9 m: Fig. 2c), but was influenced by matrix urbani-
zation (mean difference 47.6 m: P=0.018: Fig. 2d).
The effect of matrix urbanization was large, and
sugar gliders travelled almost four times farther into
less urbanized matrix (65.5 ⫾60.1 m) than into
highly urbanized matrix (17.9 ⫾12.9 m). Hierarchi-
cal partitioning demonstrated matrix urbanization
independently explained 33.8% of the total variance
explained by the data (total R2=34.6%). Matrix per-
meability was not influenced by the proximity of core
range centres to reserve boundaries (F1,16 =2.272,
P=0.151). The maximum distances travelled by
gliders into the matrix (mean ⫾SD: 36.4 ⫾44.1 m,
maximum =180.0 m) were comparable with the
typical movement distances by gliders (mean ⫾SD:
67.6 ⫾23.2 m, maximum =105.2 m).
DISCUSSION
Results from this study showed that the level of urbani-
zation in the matrix (matrix urbanization) had an
influence on matrix permeability, whereas the level of
tree cover in the matrix (matrix tree cover) did not.
These findings have important implications for con-
servation plans aimed at maintaining or improving
connectivity of gliding mammal populations in urban
areas. They suggest that low contrast urban matrix is
utilized by gliders, and therefore could be utilized to
facilitate functional connectivity in urban landscapes.
The findings also suggest that structural connectivity
is not necessarily the only factor driving matrix per-
meability for gliders. Although we did not measure
dispersal behaviour directly – rather behaviours that
Ta b l e 2 . General linear models of the main effects and interaction effects of matrix urbanization and tree cover on (1)
boundary crossing (number of locations in the matrix divided by total number of locations per glider); and (2) matrix
permeability (maximum distance travelled by gliders into matrix per glider) of individual gliders
SS df MS FPPartial eta squared
1. Boundary crossing (rate of movement)
TREE 0.038 1 0.038 1.109 0.306 0.058
URBAN 0.007 1 0.007 0.217 0.647 0.012
TREE * URBAN 0.055 1 0.055 1.587 0.224 0.081
Residual 0.622 18 0.035 R2=0.140
2. Matrix permeability (maximum distance)
TREE 0.507 1 0.507 0.802 0.386 0.054
URBAN 4.514 1 4.514 7.138 0.018 0.338
TREE * URBAN 0.023 1 0.023 0.037 0.851 0.003
Residual 8.853 14 0.632 R2=0.346
Fig. 2. Main effects of matrix urbanization and tree cover on glider boundary crossing (number of locations within matrix
divided by total number of locations per glider) and matrix permeability (maximum distance travelled into matrix per glider).
Bars show means ⫾95% confidence interval.
PERMEABILITY OF URBAN MATRIX TO GLIDERS 613
© 2012 The Authors doi:10.1111/aec.12006
Austral Ecology © 2012 Ecological Society of Australia
are precursory to dispersal (Calabrese & Fagan 2004;
Baguette & van Dyck 2007) – we can assume that any
factor influencing the movement of gliders in the
matrix is also likely to affect the resistance of the
matrix to dispersal. Our findings therefore support
the suggestion that the conservation value of low con-
trast urban matrix for gliding mammals should not be
underestimated (Brearley et al. 2010), and are highly
relevant where population isolation is an immediate
threat to urban gliders (Goldingay et al. 2006).
Matrix vegetation is known to have a positive effect
on the occupancy of mammals within urban reserves,
while matrix urbanization has a negative effect (Brady
et al. 2011a,b). However, analyses of spatial patterns of
species occurrence are typically unable to determine
the influence of behavioural inhibitors (e.g. Harris &
Reed 2002; Taylor et al. 2006; Hodgson et al. 2011).
By observing the independent and interactive effects of
matrix urbanization and tree cover on individual
movement behaviour, our findings demonstrate that
gliders may be inhibited from using otherwise struc-
turally connected habitat (George & Crooks 2006).
This has widespread relevance because gliding
mammals are found in many parts of the world in
habitats subject to fragmentation (Goldingay 2000),
and means that connectivity strategies focused solely
on structural connectivity may not be successful unless
they also account for confounding influences from the
matrix.
Structural connectivity is critical to matrix perme-
ability for gliding mammals that are vulnerable when
moving along the ground (e.g. Selonen & Hanski
2003; Ball & Goldingay 2008; Ritchie et al. 2009). It
follows that attempts to restore connectivity for gliders
have focused almost exclusively on maintaining struc-
tural connectivity (e.g. Ball & Goldingay 2008;Taylor
& Goldingay 2009). Gliders are particularly sensitive
to abrupt changes in landscape structure because their
movement is limited by their glide capacity, which for
Petaurus sp. is around 30–40 m (van der Ree et al.
2003). However, this glide capacity means they can
also traverse relatively open habitat containing scat-
tered trees, and where tree spacing and launch-tree
height allow (Selonen & Hanski 2003; van der Ree
et al. 2003). Per cent tree cover does not reflect inter-
tree distance or tree height, which might explain why
we found no difference between glider use of matrix
with high (43 ⫾6% cover) or low tree cover (20 ⫾3%
cover). We made the assumption that proportional
matrix tree cover in a given area surrounding reserves
was indicative of structural connectivity.Yet tree cover
within the matrix is heterogeneous, and even matrix
with low tree cover can be structurally connected at
fine spatial scales; as evidenced by a glider we observed
180 m from the forest reserve in matrix with low tree
cover after traversing a single line of trees. By calcu-
lating a single tree cover measure for each buffered
reserve we smoothed any heterogeneity in tree cover
within the matrix, which may explain why we failed to
detect an effect from tree cover. Inter-tree distance
could provide a more accurate assessment of structural
connectivity for gliders, but it is not routinely mapped
and therefore is currently of limited application at the
spatial scales used for urban planning. As vegetation
structure becomes easier to map through remote
sensing, variables such as tree height and canopy-gap
distance will be easier to assess and implement in
connectivity planning.
In this study, measures of urbanization (human,
building and road density) were highly correlated,
making an understanding of the specific drivers of
glider behaviour difficult. Nonetheless, each repre-
sented the spatial intensity of human disturbance
encountered in the matrix. Gliders may avoid urban-
ized areas, or utilize them less, for a number of reasons,
including increased mortality from vehicle collisions
(Goldingay et al. 2006), increased abundance of intro-
duced predators (e.g. domestic cats Felis catus, domes-
tic dogs Canis familiaris, red fox Vulpes vulpes Juzva &
Peeters 1992), increased disturbance from humans
(Brady et al. 2009), increased nocturnal lighting (Beier
2006) and habitat degradation (e.g. smaller tree size,
fewer tree hollows, greater incidence of introduced
plant species: Brady et al. 2009; Brearley et al. 2011a).
We must caution that, because we did not measure
survival, we do not know to what extent the urban
matrix (even where low contrast) may act as an eco-
logical trap for dispersing gliders (where mortality
is increased or survival and reproduction levels
decreased). Major roads (three lanes or more) typically
act as movement barriers to gliders because they often
surpass gliders glide capacity and/or have high associ-
ated mortality (Goldingay et al. 2006; van der Ree
2006; van der Ree et al. 2010). However, minor roads
(two lanes or less) are readily crossed by gliders where
road-side vegetation allows (Goldingay et al. 2006; van
der Ree 2006; Brearley et al. 2010; van der Ree et al.
2010). Almost all roads adjacent to reserves were
minor (only Damper Creek had roads with four lanes,
which accounted for 16% of roads at that reserve),
which suggests that road size was not confounding the
influence of matrix urbanization per se. But the cumu-
lative effect of high road density, even when roads are
theoretically traversable, may constrain glider move-
ment (Tremblay & St. Clair 2011). This means our
definition of what constitutes a hard edge for gliders
may vary under different contexts, for example, the
width of a single road does not form a hard edge, but
multiple small roads may. Brearley et al. (2011b)
found major roads elicited a strong edge response from
squirrel gliders (e.g. smaller home ranges at edges
compared with reserve interiors), but suggested that
smaller roads and residential areas could represent low
contrast edges.While we largely agree with this, we add
614 F. M. CARYL ET AL.
© 2012 The Authorsdoi:10.1111/aec.12006
Austral Ecology © 2012 Ecological Society of Australia
that while the effects of certain edges may seem small,
their cumulative effects when multiplied may constrain
animal movements (Tremblay & St. Clair 2011).
Management implications and conclusions
Land managers increasingly incorporate habitat con-
nectivity into designs of reserves and landscapes, rec-
ognizing that connected patches are likely to be more
successful at supporting viable wildlife populations
than isolated patches (Siitonen et al. 2002). Indeed,
the viability of threatened populations of gliding
mammals in Australian cities is dependent on the con-
nectedness of habitat patches (Goldingay et al. 2006).
Our results suggest that structural connectivity may
not be enough to maintain landscape connectivity for
gliders in urban areas if surrounding urbanization
levels are too high. Unless connectivity plans account
for the behavioural needs of target species, their aims
(increased dispersal and ultimately, population viabil-
ity) may not be successful.That said, our findings also
demonstrate that we should avoid basing connectivity
plans on discrete distinctions between habitat and
non-habitat and instead focus on specific properties
that drive animal movement (e.g. inter-tree gap, road
density, etc.). Residential housing is usually the domi-
nant land use within urban areas, so ignoring its inher-
ent heterogeneity is neglecting the potential for
numerous resources (residential gardens, street trees,
etc.) to be managed as movement habitat. All of our
study areas were surrounded by medium- to high-
density residential housing with varying levels of tree
cover and road density, yet some of these areas were
able to support glider movement without construction
of wildlife corridors. Matrix management could be
used to reduce edge-contrast around existing parks
and habitat linkages for gliders in urban areas by lim-
iting human-induced disturbance (e.g. capped housing
density, pet control). Maintaining tall trees a relatively
short distance apart (<20 m: Ball & Goldingay 2008)
along residential streets and in backyards would likely
increase their suitability for gliders.
ACKNOWLEDGEMENTS
We thank numerous volunteers who assisted with the
trapping and radio-tracking of sugar gliders, and local
residents who allowed us access to their properties.
Parks Victoria, Knox Environment Society, Monash
City Council and The Baker Foundation provided
funds for this research. Fieldwork was approved by
the University of Melbourne Animal Ethics Commit-
tee (05021) and Flora and Fauna Permit number
10004540.
REFERENCES
Australian Bureau of Statistics (2003) 2001 Census of Population
and Housing (ArcView Shapefile). Australian Bureau of Sta-
tistics, Canberra.
Australian Bureau of Statistics (2012) Regional Population
Growth,Australia, 2009–10 (Catalogue Number 3218.0). Aus-
tralian Bureau of Statistics, Canberra.
Baguette M. & van Dyck H. (2007) Landscape connectivity and
animal behaviour: functional grain as a key determinant for
dispersal. Landscape Ecol. 22, 1117–29.
Ball T. M. & Goldingay R. L. (2008) Can wooden poles be used
to reconnect habitat for a gliding mammal? Landscape Urban
Plan 87, 140–6.
Beier P. (2006) Effects of artificial night lighting on terrestrial
mammals. In: Ecological Consequences of Artificial Night
Lighting (eds C. Rich & T. Longcore) pp. 19–42. Island
Press, Washington.
Beier P. & Noss R. F. (1998) Do habitat corridors provide
connectivity? Conserv. Biol. 12, 1241–52.
Brady M. J., McAlpine C. A., Miller C. J., Possingham H. P. &
Baxter G. S. (2011b) Mammal responses to matrix devel-
opment intensity. Austral Ecol. 36, 35–44.
Brady M. J., McAlpine C. A., Possingham H. P., Miller C. J. &
Baxter G. S. (2009) Habitat attributes of landscape mosaics
along a gradient of matrix development intensity: matrix
management matters. Landscape Ecol. 24, 879–91.
Brady M. J., McAlpine C. A., Possingham H. P., Miller C. J. &
Baxter G. S. (2011a) Matrix is important for mammals in
landscapes with small amounts of native forest habitat.
Landscape Ecol. 26, 617–28.
Brearley G., Bradley A., Bell S. & McAlpine C. (2010) Influence
of contrasting edges on the abundance of arboreal
mammals: a study of squirrel gliders (Petaurus norfolcensis)in
southeast Queensland, Australia. Biol. Conser v. 143, 60–71.
Brearley G., Bradley A., Bell S. & McAlpine C. (2011a) Change
in habitat resources and structure near urban edges and its
influence on the squirrel glider (Petaurus norfolcensis)in
southeast Queensland, Australia. Austral Ecol. 36, 425–32.
Brearley G., McApline C., Bell S. & Bradley A. (2011b) Squirrel
glider home ranges near urban edges in eastern Australia.
J. Zool. 285, 256–65.
Calabrese J. M. & Fagan W. F. (2004) A comparison-shopper’s
guide to connectivity metrics. Front.Ecol. Environ. 2, 529–36.
Department of Sustainability and Environment (2007)
Corporate Geospatial Data Library, Melbourne, Austra-
lia. [Cited 3 September 2007.] Available from URL:
http://www.giconnections.vic.gov.au/content/vicgdd/record/
ANZVI0803002395.htm
Fahrig L. (2003) Effects of habitat fragmentation on
biodiversity. Annu. Rev. Ecol. Syst. 34, 487–515.
Forman R. T. T. (1995) Land Mosaics: The Ecology of Landscape
and Regions. Cambridge University Press, Cambridge.
George S. L. & Crooks K. R. (2006) Recreation and large
mammal activity in an urban nature reserve. Biol. Conserv.
133, 107–77.
Goldingay R. L. (2000) Gliding mammals of the world: diversity
and ecological requirements. In: Biology of Gliding Mammals
(eds R. L. Goldingay & J. S. Schiebe) pp. 376–400. Filander
Press, Furth.
Goldingay R. L. & Jackson S. (2004) The Biology of Australian
Possums and Gliders. Surrey Beatty and Sons PTY LTD,
Chipping Norton.
Goldingay R. L. & Sharpe D. J. (2004) How do we conserve the
squirrel glider in Brisbane’s urban matrix? In: Conservation
PERMEABILITY OF URBAN MATRIX TO GLIDERS 615
© 2012 The Authors doi:10.1111/aec.12006
Austral Ecology © 2012 Ecological Society of Australia
of Australia’s Forest Fauna, 2nd edn. (ed. D. Lunney)
pp. 663–77. Royal Zoological Society of New South Wales,
Mosman.
Goldingay R. L., Sharpe D. J., Beyer G. L. & Dobson M. (2006)
Using ecological studies to understand the conservation
needs of the squirrel glider in Brisbane’s urban forest-
remnants. Aust. Mammal. 28, 173–86.
Goldingay R. L. & Taylor B. D. (2009) Gliding performance and
its relevance to gap crossing by the squirrel glider (Petaurus
norfolcensis). Aust. J. Zool. 57, 99–104.
Haddad N. M. (1999) Corridor use predicted from behaviours
at habitat boundaries. Am. Nat. 153, 215–27.
Hahs A. K. & McDonnell M. J. (2006) Selecting independent
measures to quantify Melbourne’s urban-rural gradient.
Landscape Urban Plan 78, 435–48.
Harris R. J. & Reed J. M. (2002) Behavioural barriers to non-
migratory movement of birds. Ann. Zool. Fenn. 39, 275–90.
Harris S., Cresswell W. J., Forde P. G., Trewhella W. J., Woollard
T. & Wray S. (1990) Home-range analysis using radiotrack-
ing data – a review of problems and techniques particularly
as applied to the study of mammals. Mamm. Rev. 20,
97–123.
Hodgson P., French K. & Major R. E. (2011) Avian movement
across abrupt ecological edges: differential responses to
housing density in an urban matrix. Landscape Urban Plan
79, 266–72.
Juzva K. & Peeters P. J. (1992) A preliminary report on the status
of Sugar Gliders Petaurus breviceps (Waterhouse) reintro-
duced into Blackburn Lake Sanctuary, Victoria. Vic. Nat.
109, 168–72.
Land Victoria (2001) Vicmap Transport (ArcView Shapefile)
1:25,000. Department of Sustainability and Environment,
Melbourne.
Leary E. J. & McDonnell M. J. (2001) The patterns of public open
space in metropolitan Melbourne. Aust. Parks Leis. 4, 34–6.
Lindenmayer D. B., Lacy R. C., Tyndale-Bascoe H., Taylor A.
C., Viggers K. L. & Pope M. L. (1999) The response of
arboreal marsupials to landscape context: a large-scale frag-
mentation study. Ecol. Appl. 9, 594–611.
Noss R. F. (2007) Focal species for determining connectivity
requirements in conservation planning. In: Managing and
Designing Landscape for Conservation: Moving from Perspec-
tives to Principles (eds D. B. Lindenmayer & R. J. Hobbs)
pp. 263–79. Blackwell Publishing, Oxford.
Oates A. & Taranto M. (2001) Vegetation Mapping of the Port
Phillip and Westernport Region. Arthur Rylah Institute for
Environmental Research, Melbourne.
Prevedello J. A. & Vieira M. V. (2010) Does the type of matrix
matter? A quantitative review of the evidence. Biodivers.
Conserv. 19, 1205–23.
Quin D. G. (1995) Population ecology of the squirrel glider
(Petaurus norfolcensis) and the sugar glider (P. breviceps) (Mar-
supialia: Petauridae) at Limeburners Creek, on the central
north coast of New South Wales. Wildl. Res. 22, 471–505.
Ricketts T. H. (2001) The matrix matters: effective isolation in
fragmented landscapes. Am. Nat. 58, 87–99.
Ritchie L. E., Betts M. G., Forbes G. & Vernes K. (2009) Effects
of landscape composition and configuration on northern
flying squirrels in a forest mosaic. For. Ecol. Manage. 257,
1920–9.
Rodgers A. R. & Carr A. P. (1998) Home Range Extension (HRE)
for ArcView. Ontario Ministry of Natural Resources, Centre
for Northern Forest Ecosystem Research, Thunder Bay,
ON.
Saunders D. A. (2007) Connectivity, corridors and stepping
stones. In: Managing and Designing Landscape for Conserva-
tion: Moving from Perspectives to Principles (eds D. B. Linden-
mayer & R. J. Hobbs) pp. 280–9. Blackwell Publishing,
Oxford.
Schmiegelow F. K. A. (2007) Corridors, connectivity and bio-
logical conservation. In: Managing and Designing Landscape
for Conservation: Moving from Perspectives to Principles (eds D.
B. Lindenmayer & R. J. Hobbs) pp. 249–62. Blackwell Pub-
lishing, Oxford.
Selonen V. & Hanski I. K. (2003) Young flying squirrels (Pter-
omus volans) dispersing in fragmented forests. Behav. Ecol.
15, 564–71.
Sharpe D. J. & Goldingay R. L. (2007) Home range of the
Australian squirrel glider Petaurus norfolcensis (Diproto-
dontia). J. Mammal. 88, 1515–22.
Siitonen P., Tanskanen A. & Lehtinen A. (2002) Method for
selection of old-forest reserves. Conserv. Biol. 16, 1398–408.
Simberloff D. & Cox J. (1987) Consequences and costs of con-
servation corridors. Conserv. Biol. 1, 63–71.
Simberloff D., Farr J. A., Cox J. & Mehlman D. W. (1992)
Movement corridors: conservation bargains or poor invest-
ments? Conserv. Biol. 6, 493–504.
Stamps J. A., Buechner M. & Krishnan V. V. (1987) The effects
of edge permeability and habitat geometry on emigration
from patches of habitat. Am. Nat. 129, 533–52.
Suckling G. C. (1984) Population ecology of the sugar glider,
Petaurus breviceps, in a system of fragmented habitats. Wildl.
Res. 11, 49–75.
Taylor B. D. & Goldingay R. L. (2009) Can road-crossing struc-
tures improve population viability of an urban gliding
mammal? Ecol. Soc. 14, 13–32.
Taylor P., Fahrig L. & With K. A. (2006) Landscape connectiv-
ity: back to the basics. In: Connectivity Conservation (eds K.
Crooks & M. A. Sanjayan) pp. 29–43.Cambridge University
Press, Cambridge.
Tischendorf L. & Fahrig L. (2000) How should we measure
landscape connectivity? Landscape Ecol. 15, 633–41.
Tremblay M. A. & St. Clair C. (2011) Permeability of a hetero-
geneous urban landscape to the movements of forest
songbirds. J. Appl. Ecol. 48, 679–88.
van der Ree R. (2002) The population ecology of the Squirrel
Glider Petaurus norfolcensis within a network of remnant
linear habitats. Wildl. Res. 29, 329–40.
van der Ree R. (2006) Road upgrade in Victoria a filter to the
movement of the endangered squirrel glider (Petaurus nor-
folcensis): results of a pilot study. Ecol. Manage. Restor. 7,
226–8.
van der Ree R. & Bennett A. F. (2003) Home range of the
squirrel glider (Petaurus norfolcensis) in a network of remnant
linear habitats. J. Zool. 259, 327–36.
van der Ree R., Bennett A. F. & Gilmore D. C. (2003) Gap-
crossing by gliding marsupials: thresholds for use of isolated
woodland patches in an agricultural landscape. Biol.
Conserv. 115, 241–9.
van der Ree R., Cesarini S., Sunnucks P., Moore J. L. & Taylor
A. C. (2010) Large gaps in canopy reduce road crossing by
a gliding mammal. Ecol. Soc. 15, 35.
Watling J. I., Nowakowski A. J., Donnelly M. A. & Orrock J. L.
(2011) Meta-analysis reveals the importance of matrix com-
position for animals in fragmented habitat. Global Ecol.
Biogeogr. 20, 209–17.
White G. C. & Garrott R. A. (1990) Analysis of Wildlife Radio-
tracking Data. Academic Press, San Diego, CA.
616 F. M. CARYL ET AL.
© 2012 The Authorsdoi:10.1111/aec.12006
Austral Ecology © 2012 Ecological Society of Australia
