Edge effects on plant communities along power line clearings

Abstract

Power line clearings are edge‐creating disturbances in landscapes world‐wide, but there have been few studies on their bordering vegetation. Our aim was to quantify edge effects on plant communities along such clearings in Norway and to identify factors that influence these edge effects.
We surveyed understorey plant communities on either side of the power line clearing–forest edge at 51 sites, along four parallel transects at each site. Each transect had four plots located, respectively, in the clearing centre, clearing edge, forest edge and forest. We quantified the magnitude of edge effects (MEE) on either side by comparing edges with their corresponding ‘non‐edge’ reference habitats. We also measured differences in species composition across the edge (clearing edge vs. forest edge). Habitat characteristics were sampled at plot and site level and from digital maps.
Differences in species composition were greater between clearing centres and clearing edges than between forests and forest edges. Differences in species composition across the edge increased with edge contrast and forest productivity. Edge effects on species composition into the forest were smallest along north‐facing edges, whereas those in the clearings increased with power line age.
Species richness increased slightly towards the edge in forests but decreased considerably towards the edge in clearings. The direction and MEE on either side differed among functional groups. Edge contrast and edge aspect were the prime factors influencing the MEE into forests, whereas in clearings, these were influenced principally by tree regrowth in the clearings and by forest productivity.
Synthesis and applications . Edge effects on plant communities bordering power line clearings were determined by factors that can be influenced by planners and managers. For existing power lines, management plans should differentiate between the following: (i) clearings through high conservation value forests, where edge effects into the adjacent forest should be limited; (ii) clearings that can act as replacement habitat for cultural landscape species, where maintaining open‐canopy habitats should be prioritized; and (iii) ‘business‐as‐usual’ clearings, where continuing the current practice of cutting every 5–10 years is recommended.

Full text

Edge effects on plant communities along power line
clearings
Katrine Eldegard*, Ørjan Totland and Stein R. Moe
Department of Ecology and Natural Resource Management, Faculty of Environmental Science and Technology,
Norwegian University of Life Sciences, PO Box 5003, NO-1432

As, Norway
Summary
1. Power line clearings are edge-creating disturbances in landscapes world-wide, but there
have been few studies on their bordering vegetation. Our aim was to quantify edge effects on
plant communities along such clearings in Norway and to identify factors that influence these
edge effects.
2. We surveyed understorey plant communities on either side of the power line clearing–for-
est edge at 51 sites, along four parallel transects at each site. Each transect had four plots
located, respectively, in the clearing centre, clearing edge, forest edge and forest. We quanti-
fied the magnitude of edge effects (MEE) on either side by comparing edges with their corre-
sponding ‘non-edge’ reference habitats. We also measured differences in species composition
across the edge (clearing edge vs. forest edge). Habitat characteristics were sampled at plot
and site level and from digital maps.
3. Differences in species composition were greater between clearing centres and clearing
edges than between forests and forest edges. Differences in species composition across the
edge increased with edge contrast and forest productivity. Edge effects on species composition
into the forest were smallest along north-facing edges, whereas those in the clearings increased
with power line age.
4. Species richness increased slightly towards the edge in forests but decreased considerably
towards the edge in clearings. The direction and MEE on either side differed among func-
tional groups. Edge contrast and edge aspect were the prime factors influencing the MEE into
forests, whereas in clearings, these were influenced principally by tree regrowth in the clear-
ings and by forest productivity.
5. Synthesis and applications. Edge effects on plant communities bordering power line clear-
ings were determined by factors that can be influenced by planners and managers. For exist-
ing power lines, management plans should differentiate between the following: (i) clearings
through high conservation value forests, where edge effects into the adjacent forest should be
limited; (ii) clearings that can act as replacement habitat for cultural landscape species, where
maintaining open-canopy habitats should be prioritized; and (iii) ‘business-as-usual’ clearings,
where continuing the current practice of cutting every 5–10 years is recommended.
Key-words: edge contrast, edge influence, forest influence, habitat contrast, plant community
properties, plant functional groups, species composition, species diversity, species richness
Introduction
One of the potentially negative results of human-induced
fragmentation is the artificial edge created between two
habitats (Lindenmayer & Fischer 2006). Such edges are
often abrupt, producing changes in diversity and the
structural and functional complexity of plant communities
through the interplay of direct human disturbance and
indirect changes in abiotic and biotic conditions (Harper
et al. 2005; Ewers & Didham 2006a; Braithwaite & Mallik
2012). In forests, edges typically experience increased solar
radiation, lower humidity, higher air temperature, higher
soil temperatures and increased wind speed, compared
with the forest interior (Hamberg, Lehvavirta & Kotze
2009; Braithwaite & Mallik 2012). Forest edges also often
*Correspondence author. E-mail: katrine.eldegard@nmbu.no
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society
Journal of Applied Ecology 2015, 52, 871–880 doi: 10.1111/1365-2664.12460

have higher stem densities, greater species richness and
proportionately more shade-intolerant and exotic species
(Fahrig 2003; Harper et al. 2005). Edges can therefore
have large impacts on surrounding species diversity, com-
position, community dynamics and ecosystem functioning
(Laurance et al. 2007).
Establishing and maintaining power line clearings is
one type of human activity that creates so-called through
corridors of earlier successional vegetation and long, per-
manent edges (Forman 1995). Some studies on the vegeta-
tion characteristics of power line clearings have been
conducted (Forman 1995; Dumas et al. 2003), but despite
the extensive network of such clearings in many land-
scapes (Wojcik & Buchmann 2012) and the considerable
literature on edge effects, few have attempted to quantify
edge effects on plant communities in adjacent forest habi-
tats (but see Luken, Hinton & Baker 1991; Powell &
Lindquist 2011). In contrast, edge effects on plant com-
munities along roads, another common type of ‘through
corridor’, have received more attention (Forman et al.
2003; Laurance, Goosem & Laurance 2009). Linear utility
corridors, such as roads and power line clearings, have
been thought to affect biodiversity negatively (Nekola
2012; Wojcik & Buchmann 2012). During the last decade,
however, more studies have highlighted their potential
value as habitats for several groups of animals: bees (Rus-
sell, Ikerd & Droege 2005), butterflies (Forrester, Leopold
& Hafner 2005; Berg et al. 2011), terrestrial gastropods
(Nekola 2012), and early successional birds and mammals
(Clarke & White 2008; Askins, Folsom-O’Keefe & Hardy
2012).
Changes in plant community composition and diversity
along human-induced edges have been studied primarily
at forest edges abutting clear-cuts and agricultural land
(Harper et al. 2005; Baker et al. 2013; Chabrerie et al.
2013; Pellissier et al. 2013). Power line clearings resemble
forest clear-cuts, but differ by being linear and having
greater edge-to-area ratios. Regular cutting of trees in
such clearings maintains the vegetation at an early succes-
sional stage, whereas the vegetation in clear-cuts is left to
regrow. After maintenance cutting of power line clearings,
the dead wood is usually left to decay; nutrient loss is
probably less than after timber harvest. Power line clear-
ings and clear-cuts can also differ in the effect of the edge
on adjacent forest and of the forest on the disturbed area
(‘forest influence’; Baker et al. 2013). Recent studies have
focused on using this forest influence on harvested areas
to re-establish forest biodiversity (Baker et al. 2013). In
contrast, re-establishing forest biodiversity in power line
clearings could be unfeasible, because of large contrasts in
important habitat conditions.
Edge effects are trans-boundary phenomena and are
assumed to be affected both by the degree of abiotic and
biotic contrasts between the two adjacent habitats and by
characteristics of the edge itself (Harper et al. 2005;
Barnes et al. 2014). Edge effects can also depend on attri-
butes of the surrounding landscape; diverse, structurally
complex landscapes can neutralize differences between
habitat edges and interiors (Ewers & Didham 2006b).
Theoretical models predict edge effects on both sides of
the habitat boundary, yet one-sided approaches, which
study ecological patterns and processes from an edge to
the interior of only one of the habitats, predominate
(Fonseca & Joner 2007). Two-sided approaches, focusing
on ecological patterns and processes across the whole gra-
dient from the interior of one habitat through the edge
zone to the interior of the other, are much less common
(Fonseca & Joner 2007). Even then, while studies of edge
effects on forest biodiversity commonly focus on gradients
from clear-cuts into the unlogged forest, the reciprocal
influence, of the forest on disturbed areas, is usually
ignored (Baker et al. 2013).
Power line clearings are well suited to studies of these
complex edge effects. The initial clearing of the power
line, followed by repeated cutting of the tree vegetation,
can be viewed as a quasi-experiment (Lindenmayer & Fi-
scher 2006), broadly similar to genuine large-scale experi-
ments. In this paper, we report on the edge effects of
power line clearances on community-level measures of
plant biodiversity (species composition, richness and
diversity) and on easily recognized functional groups of
plants. Our aim was (i) to quantify the magnitude of edge
effects (MEE) on understorey plant communities on either
side of forest–power line edges; (ii) to measure difference
in species composition across the edge and (iii) to identify
factors potentially determining these edge effects. We
included factors that planners and managers can influ-
ence. We hypothesized that the MEE on species composi-
tion, richness and diversity would be greater in the
clearings than on the forest side of the edge, because of
steeper gradients in environmental conditions (Belyea &
Lancaster 1999). We also expect the direction and MEE
to vary among functional groups in the plant community
and that the attributes of the edges, power line clearing
and forest will influence these effects.
Materials and methods
STUDY AREA AND SAMPLING DESIGN
The 51 study sites were haphazardly distributed across the main
power line grid in south-east Norway (Fig. 1). The sites were
located between latitudes 58°–61°N and longitudes 8–11°Eat
25–1055 m a.s.l. Average temperatures in January range from
102to26°C, rising in May–July to 79–155°C. Monthly
precipitation averages 47–103 mm. All the sites were situated in
boreal forests intersected by a power line, with varying propor-
tions of the main tree species: Norway spruce Picea abies, Scots
pine Pinus sylvestris and birch Betula spp. At all but one site (a
forest reserve), the adjacent forest was managed for timber har-
vest. The clearings varied in width and age (Table 1), but were
subject to the same management regime: cutting of all woody
vegetation in the clearing every 5–10 years (no chemicals used).
At each site, we deployed four parallel transects 50 m apart,
each consisting of four plots, located, respectively, in the clearing
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
872 K. Eldegard, Ø. Totland & S. R. Moe

centre (the clearing reference point), clearing edge, forest edge
and forest (the forest reference point) (Fig. 1). The location of
the edge was determined as the edge of the canopy of the adja-
cent forest. The edge plots on either side of the forest–power line
edge were located median 3 m apart (interquartile range: 2–5).
The two plots (reference and edge) on either side of the edge con-
stituted a plot-pair. Each plot measured 4 95 m. We placed five
1-m
2
subplots along the centre line of each plot. We also sampled
further into the forest, but there was little difference between for-
est reference and interior plots (see Table S1, Supporting Infor-
mation).
DATA COLLECTION
Field data on understorey vegetation and habitat characteristics
were collected at 20 sites in 2009 and 31 sites in 2010. We visually
estimated total cover of all vascular plant species in five 1-m
2
subplots within each of the 20 plots at every site (Fig. 1). If a
species was present in a subplot, but had <1% cover, it was
recorded as 1%. Small trees <1 m height were recorded as under-
storey shrubs. We also recorded the total cover of mosses and
lichens (not to species level) within each subplot.
In all 20 plots at each site, we recorded species, height and
crown width (by visual estimation) of all trees >1 m tall. We also
recorded basal area (relascope sum) from the centre of each plot.
Forest productivity potential (hereafter termed forest productiv-
ity) for the forest plots was determined from the dominant tree
species, soil depth, terrain slope, height above sea level, northern
latitude and the vegetation type that dominated an area of 100
m
2
(radius 564 m) around the centre of the plot (Nilsen & Lars-
son 1992).
The power line grid owner (Statnett) provided data on the year
of establishment of the power line clearings. For each site, we
extracted data on edge aspect (degrees deviation from the south)
and clearing width from aerial photos (http://www.norgeibil-
der.no/). To assess the degree of forest fragmentation, we
used ArcGIS (ESRI 2011) and Ar5 digital maps (http://www.
skogoglandskap.no/temaer/ar5) to measure the length of forest
edges at increasing radii around each site (Table 1).
DATA ANALYSES
We calculated compositional dissimilarity from the difference in
species composition between reference and edge plots within each
plot-pair (Fig. 1) and between edge plots on either side of the
edge (Chao et al. 2008, http://chao.stat.nthu.edu.tw//softwa-
reCE.html). Input data were the summed species abundances
from the five subplots within each plot. To compare edge effects
on either side of the edge on species composition, we fitted a gen-
eralized linear mixed model (GLMM, l
2
9(1 l
2
) variance
structure, logit link) with side as fixed effect and plot-pair nested
within site as random effects (SAS 2011). We also quantified the
difference in species composition between reference and edge
plots as the difference in plot scores generated in a global non-
metric multidimensional scaling (GNMDS) ordination of the vas-
cular plant data (see Fig. S1).
Plant data from the five 1-m
2
subplots within each 4 95m
plot were used to calculate total species richness and Shannon
diversity as measures of richness and diversity per plot. We also
calculated species richness and total cover of six different func-
tional groups –shrubs, dwarf shrubs, graminoids, ferns and
allies, shade-tolerant forbs and shade-intolerant forbs –and the
total cover of mosses and lichens. Shade-tolerance scores for forb
species were derived from Hill et al. (1999). We categorized spe-
cies with Light (L) values of 1–5 as shade tolerant (30 species)
and those with L values of 6–9 (96 species) as shade intolerant.
To compare the values of all these response variables among our
four ‘habitats’ –clearing centre, clearing edge, forest edge and
forest –we fitted GLMMs (SAS 2011) for each response variable,
with habitat as the fixed effect and site and plot-pair nested
within site as random effects. For the richness response variables,
we fitted models with a Poisson distribution and a log link func-
tion. The total cover variables were converted to proportions and
modelled with a [l
2
9(1 l
2
)] variance structure and logit link.
(b)
(a)
Fig. 1. (a) The 51 study sites and (b) sche-
matic illustration of a study site. In (a),
bubble size reflects the relative total plant
species richness. C =clearing (reference),
Ce =clearing edge, Fe =forest edge and
F=forest (reference). Understorey cover
data were collected within 1-m
2
subplots
placed along the centre line of the
495 m plots.
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
Edge effects along power line clearings 873

For diversity, we fitted a model with normal distribution and
identity link.
The MEE on total species richness and diversity of vascular
plants was calculated as [(edge plot reference plot)/(edge
plot +reference plot)] on either side of the edge (Harper et al.
2005). A positive edge effect on, for example, species richness
means that richness was higher in edge plots than in reference
plots, whereas a negative edge effect means the opposite. We used
the same formula to calculate the MEE on the richness and cover
of each functional group.
To quantify edge contrast, we calculated the absolute differ-
ence in tree-variable measures between edge plots on each side
of the edge. We used individual measures of trees in the clear-
ing centre plots to quantify tree regrowth in the clearings. To
deal with colinearity among the various measures of (i) edge
contrast; (ii) tree regrowth and (iii) amount of forest edge in
different buffer zones, we performed separate principal compo-
nents analysis (PCA) on each group of variables. We used PCA
axis 1 scores as explanatory variables in the analyses of factors
that influence edge effects (Table 1). The PCA-derived explana-
tory variables were standardized to zero skewness and rescaled
to the range 0–1 before analysis (Økland, Økland & Rydgren
2001).
The potential influence of the explanatory variables on species
composition, both across the edge (clearing edge vs. forest edge)
and on either side of the edge (forest edge vs. forest reference,
clearing edge vs. clearing reference), was analysed by fitting indi-
vidual linear mixed effects models (Pinheiro et al. 2014).
Similarly, the potential influence of the explanatory variables on
the MEE on total species richness, diversity and total cover of
each functional group was analysed for each side (forest, clearing)
separately. We fitted individual models for all the response vari-
ables and candidate explanatory variables, with plot-pair nested
within site as random effects.
Results
EFFECTS ACROSS AND ON EITHER SIDE OF THE EDGE
The largest difference in species composition occurred
between the clearing centre and the forest, whereas the
smallest difference occurred between plots on either side
of the edge (Fig. 2a,b). Edge effects on species composi-
tion were substantial on both sides of the edge (Fig. 2a,
b), but compositional dissimilarity between the centre and
the edge in the clearings was on average 20% higher than
between the forest and the forest edge (F
1,405
=196,
P<00001). There was substantial among-site and
among-plot-pairs variation in compositional dissimilarity
(likelihood ratio test; v2
2=1008, P<00001).
Species richness decreased along the gradient from the
clearing centre to the forest, most steeply from the clear-
Table 1. Explanatory variables included in the analyses of edge effects across and on either side of the edge along power line clearings.
Measures of the (i) tree layer vegetation (regrowth) in the clearings; (ii) edge contrast between clearings and adjacent forests (clearing
edge vs. forest edge) and (iii) forest edges in the surrounding landscape were combined in separate principal component analyses (PCA)
Variables in analyses Variables included in PCA
Correlation (r)
with PCA axis 1
Spatial
scale
Measured values
(range)
Power line clearing attributes
Width Site 25–73 m
Age (years since first clearing) Site 24–81 years
Regrowth of trees in clearing centre
Relascope sum 06 Plot 0–36 m
2
ha
1
No. trees >1m 08 Plot 0–66 trees
No. conifers >1m 05 Plot 0–25 trees
No. deciduous trees >1m 07 Plot 0–61 trees
No. spruce trees >1m 05 Plot 0–10 trees
No. pine trees >1m 02 Plot 0–25 trees
Mean tree height of trees >1m 08 Plot 0–7m
Max tree height 08 Plot 0–8m
Sum crown width of trees >1m 08 Plot 0–479m
Edge attributes
Edge aspect (deviation from the south) Site 4–166 degrees
Edge contrast
Average height of trees in edge 06 Site 35–12 m
Difference in mean tree height 08 Plot 0–12 m
Difference in max tree height 08 Plot 0–18 m
Difference in sum crown width 07 Plot 0–35 m
Difference in relascope sum 06 Plot 0–23 m
2
ha
1
Forest attributes
Forest productivity index Plot 59–231
Forest edges in the landscape
Within 150 m radius 06 Site 0–966 m
Within 300 m radius 08 Site 0–3082 m
Within 500 m radius 09 Site 0–6758 m
Within 1000 m radius 08 Site 373–26 088 m
Within 2000 m radius 07 Site 6567–94 826 m
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
874 K. Eldegard, Ø. Totland & S. R. Moe

ing centre to its edge (Fig. 3). Only one invasive species
(Epilobium ciliatum) was found in one (clearing centre)
plot at one site. The edge effect into the forest occurred
as a slight increase in species richness at the forest edge
(compared with the forest), whereas the forest influence
on species richness in the clearing was negative (Fig. 4).
For diversity, there was a similar pattern of forest influ-
ence in the clearing, whereas we found no edge effect into
the adjacent forest (Figs 3, 4).
Richness and cover of shrubs, shade-intolerant forbs
and graminoids decreased curvilinearly (linearly in the
case of graminoid richness) from the clearing centre to the
forest. The decrease was greatest from the clearing centre
to the clearing edge, whereas there was a positive edge
effect in forest, with increased richness and cover in the
forest edge (Figs 3, 4). For dwarf shrubs and lichens, we
found a humped pattern with the highest richness and
cover at the edge on both sides of the edge. For ferns and
allies, the edge plots on both sides had the lowest richness
and cover. For shade-tolerant forbs, the edge effects on
richness and cover on both sides appeared weak. Moss
cover declined substantially from the forest to the forest
edge, whereas it was similar in the forest edge, clearing
edge and clearing habitats.
FACTORS INFLUENCING EDGE EFFECTS
The first axis of the PCA of the nine measures of trees in
the clearing centres accounted for 51% of total variation
in the data and correlated strongly with most of the indi-
vidual measures of trees (Table 1). Thus, increasing values
of this PCA axis likely represents increasing regrowth of
trees in the clearing. The first axis of the PCA of the five
measures of contrast in tree variables between edge plots
on each side of the edge also accounted for 51% of the
total variation and correlated strongly with all the individ-
ual measures of edge contrast (Table 1). Increasing values
of this PCA axis likely represent increasing edge contrast.
PCA axis 1 of the measures of length of forest edge in the
five buffer zones around each site accounted for 68% of
total variation and was highly correlated with all the indi-
vidual measures (Table 1). Thus, increasing values of this
PCA axis likely represent more forest edge in the sur-
rounding landscape.
The difference in species composition (compositional
dissimilarity) between adjacent plots on either side of the
edge increased with forest productivity (t
1,152
=23,
P=0022, Fig. 2c) and edge contrast (t
1,152
=21,
P=0041, Fig. 2d).
The difference in species composition between forest
and forest edge diminished as the edges increasingly faced
north (Fig. 5a, t
1,49
=20, P=0049, Table S2a). For
mosses and shade-tolerant forbs, the overall negative edge
effects declined on more northward-facing edges (Fig. 4,
t
1,49
=37, P<00001, t
1,49
=22, P=0032). As edge
contrast increased, the overall positive edge effect on
dwarf shrubs increased (t
1,152
=26, P=0010), but on
lichens, it declined (t
1,152
=21, P=00037). The nega-
tive edge effect on mosses intensified (t
1,152
=28,
P=00051). Greater amounts of forest edge in the sur-
rounding landscape reduced the overall positive effect of
edges on understorey shrubs (t
1,49
=28, P<00001).
Compostional dissimilarity
0·00 0·50 1·00
Forest influence on
clearings (Ce vs.C)
Dissimilarity across
the edge (Ce vs.Fe)
Edge effects into
forest (Fe vs.F)
Dissimilarity between
clearings and forest
(C vs.F)
(a)
–0·03 −0·02 −0·01 0·00 0·01 0·02 0·0
3
−0·06
−0·04
−0·02
0·00
0·02
0·04
0·06
GNMDS1 plot scores
GNMDS2 plot scores
(b)
●
●
●
●
C
Ce
Fe
F
0·0 0·2 0·4 0·6 0·8 1·0
0·0
0·2
0·4
0·6
0·8
1·0
Forest productivity
Compositional dissimilarity (Ce vs.Fe)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●●
●
●
●
(c)
0·0 0·2 0·4 0·6 0·8 1·
0
0·0
0·2
0·4
0·6
0·8
1·0
Edge contrast
Compositional dissimilarity (Ce vs.Fe)
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
(d)
Fig. 2. Differences in species composition
from the centre of the clearings, across the
edge and into the adjacent forest: (a) com-
positional dissimilarity (0 =identical,
1=no species shared), bars and whiskers
are observed means 1 SE; (b) displace-
ment along global non-metric multidimen-
sional scaling ordination axes 1 and 2
(arrows show the average change in plot
scores, see Fig. S2); (c) and (d): estimated
relationship (SE) between compositional
dissimilarity and forest productivity and
edge contrast, respectively. Dots are
observed values.
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
Edge effects along power line clearings 875

The difference in species composition between forest edge
and the forest tended to increase with corridor width and
thus with the distance between the forest edge and forest
plots (t
1,49
=18, P=0081).
The age of the power line influenced the contrast in
species composition and moss cover between the edge
and centre of clearings. The difference in species compo-
sition increased with age (t
1,49
=22, P=0032, Fig. 5b,
Table S2b), whereas the overall positive effect on moss
cover, with increased cover at the edge relative to the
centre, lessened with age (t
1,49
=24, P=0020).
Increasing productivity of the adjacent forest reduced the
overall negative effects of edges on species richness,
diversity, graminoids and shade-tolerant forbs (t
(rich)
1,152
=23, P=0021, t
(div)1,152
=20, P=0043, t
(gram)
1,152
=22, P=0029, t
(forb)1,152
=22, P=0028, Figs 4,
5c). The difference in vascular plant species composition
between the edge and centre of the clearings declined as
trees regrew (t
1,152
=19, P=0056). The overall nega-
tive edge effect on species richness, diversity and the
cover of shrubs, and ferns and allies were accentuated
(t
(rich)1,152
=21, P=0034, t
(div)1,152
=22, P=0033,
t
(shrub)1,152
=21, P=0034, t
(fern)1,152
=25,
P=0014, Figs 4, 5d), whereas on lichen cover, the
effect became increasingly positive (t
1,152
=36,
P=00004). As edge contrast increased, the overall posi-
tive effects on the cover of dwarf shrub also increased
(t
1,152
=31, P=00020). Increasing amounts of forest
edge in the surrounding landscape increased the overall
positive effects on the cover of lichens (t
1,49
=19,
P=0057) and reduced it on the cover of dwarf shrubs
(t
1,49
=25, P=0014).
Richness/diversity
Number of species
CCeFeF
8
9
10
11
12
13
14
●
●
●
●
●
●
●
●
Vascular plants
Number of species
CCeFeF
1·0
1·5
2·0
2·5
3·0
●
●
●
●
●
●
●
●
Shrubs
CCeFeF
2·2
2·4
2·6
2·8
3·0
●
●
●
●
●
●
●
●
Dwarf shrubs
CCeFeF
0·0
0·2
0·4
0·6
0·8
1·0
●
●
●
●
●
●
●
●
Shade−intolerant forbs
Shannon entropy
CCeFeF
1·4
1·6
1·8
2·0
●
●
●
●
●
●●
●
Vascular plants
Number of species
CCeFeF
0·0
0·1
0·2
0·3
0·4
0·5
0·6
●
●
●
●
●
●
●
●
Ferns and allies
CCeFeF
1·0
1·5
2·0
2·5
3·0
●
●
●
●
●
●
●
●
Graminoids
CCeFeF
1·0
1·5
2·0
2·5
●
●
●
●
●
●●
●
Shade−tolerant forbs
Cover
Total cover (proportion)
CCeFeF
0·00
0·02
0·04
0·06
0·08
●
●
●
●
●
●
●
●
Lichens
CCeFeF
0·00
0·01
0·02
0·03
0·04
0·05
●
●
●
●●
●
●
●
Shrubs
CCeFeF
0·00
0·05
0·10
0·15
0·20
0·25
●
●
●
●●
●
●
●
Dwarf shrubs
CCeFeF
0·00
0·01
0·02
0·03
0·04
0·05
●
●
●
●●
●
●
●
Shade−intolerant forbs
Total cover (proportion)
CCeFeF
0·0
0·1
0·2
0·3
0·4
0·5
●
●●
●●
●
●
●
Mosses
CCeFeF
0·00
0·02
0·04
0·06
0·08
0·10
●
●
●
●●
●
●
●
Ferns and allies
CCeFeF
0·00
0·05
0·10
0·15
0·20
0·25
●
●
●
●
●
●●
●
Graminoids
CCeFeF
0·00
0·01
0·02
0·03
0·04
0·05
●
●
●
●
●
●●
●
Shade−tolerant forbs
Fig. 3. Observed mean (1 SE) values of response variables from the centre of the clearings (C), across the edge (Ce =clearing edge,
Fe =forest edge) and into the adjacent forest (F). Asterisks indicate significance levels (***P<0001, *P<005, 9P<010) for pairwise
comparisons of habitat means, based on generalized linear mixed models.
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
876 K. Eldegard, Ø. Totland & S. R. Moe

Discussion
The MEE on species composition, total richness and
diversity, and richness and cover of most of the functional
groups was larger in the clearings, relative to the forest.
The magnitude and direction of edge effects varied among
the functional groups. We could identify management-rel-
evant factors influencing effects on either side of or across
the edge, although there was considerable variation
among sites in such effects.
EFFECTS ACROSS THE EDGE AND INTO ADJACENT
FOREST
Species richness increased towards the edge in the forest.
Richness and cover of most functional groups also
increased, although some, such as moss cover, decreased.
These changes in the plant community likely reflect
changes in abiotic conditions such as increased solar radia-
tion and lower humidity in forest near the edge, and dis-
persal of species from clearings into the forest.
Maintenance of linear clearings, which prevents natural
edge closure, can intensify edge effects (Laurance, Goosem
& Laurance 2009). Powell & Lindquist (2011) and Luken,
Hinton & Baker (1991) found that species composition,
richness and diversity of trees did not differ between forest
edges and interiors in forests adjacent to power line clear-
ings, but their results are not directly comparable to ours,
as they did not study understorey plant communities. Bio-
tic and abiotic edge effects are of particular concern in
old-growth natural forests because they can be detrimental
to forest interior species, including many bryophytes (Lin-
denmayer & Franklin 2002). In Europe, the amount of
forests of high conservation value, typically natural old-
growth forest, is decreasing (Bengtsson et al. 2000). Thus,
throughout Europe, power line clearings commonly tran-
sect intensively managed forests that are often dense regen-
erating stands of low structural complexity and
biodiversity (Bengtsson et al. 2000). In such forests, edge
effects into forests are not necessarily detrimental, as these
forests rarely harbour vulnerable forest interior species.
The increased turnover of species across the edge with
increasing edge contrast and productivity accords with
predictions from conceptual models and empirical studies
of edge effects at human-induced forest edges (Harper
et al. 2005; McWethy, Hansen & Verschuyl 2009). For-
man et al. (2003) reviewed studies of the spread of road-
side plant species into adjacent forest and found that
most invaded only up to about 10 m from a roadside,
although the potential influence of edge contrast along
roads remains uncertain. Clarke & White (2008) recom-
mended softening abrupt edges at the interface between
forest and power line clearings, to provide habitat for
wildlife species and reduce the invasion of introduced spe-
cies into the adjacent forest.
The MEE into the forest on species composition,
mosses and shade-tolerant forbs were smallest on north-
facing edges. Our results accord with previous studies,
showing that the orientation of linear clearings and other
human-induced forest edges influences the intensity of
edge effects on plant communities (Forman et al. 2003;
Hamberg, Lehvavirta & Kotze 2009; Laurance, Goosem
& Laurance 2009). Increasing deviation from the south in
the northern hemisphere reduces edge effects on microcli-
mate, light regimes and associated biotic changes
(Hylander 2005).
We found a weak tendency for greater difference in spe-
cies composition between forest edge and forest with
Magnitude of edge effect
s
–0·4 –0·1 0·1 0·3
Biological diversity (vascular plants)
Species richness
Diversity
Functional groups richness
Shrubs
Dwarf shrubs
Graminoids
Ferns and allies
Shade−intolerant forbs
Shade−tolerant forbs
Functional groups cover
Shrubs
Dwarf shrubs
Graminoids
Ferns and allies
Shade−intolerant forbs
Shade−tolerant forbs
Mosses
Lichens
Forest influence on clearings
Edge effects into forest
Fig. 4. Magnitude of edge effects (MEE) on either side of the
edge along power line clearings. Bars and whiskers are observed
means 1 SE. The MEEs on either side were calculated as
(edge reference)/(edge +reference) (Fig. 1). Positive [negative]
MEEs mean that the richness, diversity or total cover was higher
[lower] in the edge than in the reference plots.
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
Edge effects along power line clearings 877

increasing corridor width and thus with the distance
between the edge plots and forest plots. The spatial extent
of the adjacent area with high contrast conditions can
influence the intensity of edge effects (Driscoll et al.
2013). In tropical forests, narrow (<20 m width) linear
clearings are less vulnerable to edge-related wind distur-
bance and desiccation stress than are wider clearings
(Laurance, Goosem & Laurance 2009). All our clearings
were ≥25 m wide, and thus, the distance between edge
and reference plots was a minimum 125 m (Fig. 1). Edge
effects on plant species composition and diversity typically
extend from <10 to 25 m into the adjacent forest (Harper
et al. 2005). Perhaps the widths of our power line clear-
ings and thereby the interplot distances (Table 1) did not
vary enough to make a difference.
Our finding that tree regrowth in the clearings did not
influence edge effects in the adjacent forest contradicts
some studies showing that increasing contrast in condi-
tions between two adjacent habitats (patch vs. matrix con-
trast) intensifies the edge effects on plant communities
(Jules & Shahani 2003; Chabrerie et al. 2013). Perhaps, in
our study area, the larger forest matrix overwhelms the
effects of the linear clearings embedded in it; perhaps, the
‘patch-matrix’ model is inadequate.
EDGE EFFECTS ON PLANT COMMUNITIES IN
CLEARINGS
Species richness decreased towards the edge in the clear-
ings, as did all the other responses, except for dwarf
shrubs, lichens and mosses. These edge effects on the
plant community can probably be ascribed to shading
from the nearby forest, which makes abiotic conditions
near the edge different from areas further out. The estab-
lishment of power line clearings changes abiotic condi-
tions by increasing solar energy, air temperatures and
wind speed (Pohlman, Turton & Goosem 2009). Larger
differences in abiotic conditions between the edge and the
reference ‘non-edge’ habitat could explain why the MEE
on species composition and other responses was larger in
the clearings, compared with the adjacent forest.
Although power line clearings are frequently disturbed
by maintenance cutting (Luken, Hinton & Baker 1992),
this disturbance regime likely maintains substantial resid-
ual soil function, plant propagules and a largely intact
topsoil and seed bank. Thus, the plant communities prob-
ably undergo secondary succession after disturbance
(Pickett, Cadenasso & Meiners 2013). The power line
clearings in our study were able to resist invasion by exo-
tic species. Our results accord with Burt & Rice (2009),
who found that cleared ski runs retained many ecological
similarities to reference forest and possessed greater plant
species diversity than the forest. In contrast, roads and
machine-graded ski runs create bare soil surfaces where
plant succession starts from bare ground. Plant communi-
ties in such clearings typically include many exotic and
invasive species (Forman et al. 2003; Burt & Rice 2009).
Road clearings also differ from power lines by having a
compacted surface along the centre line and more fre-
quent disturbance of the vegetation.
We observed the highest species richness and diversity
in the centre of clearings. Many of the plant species that
0 30 60 90 120 150 180
0·0
0·2
0·4
0·6
0·8
1·0
Edge aspect (deviation from the south)
Compositional dissimilarity
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
(a) Forest vs. forest edge
20 30 40 50 60 70 80 90
0·0
0·1
0·2
0·3
0·4
0·5
Years since first clearing
Change in species composition (GNMDS1)
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
(b) Centre vs. edge of clearing
0·0 0·2 0·4 0·6 0·8 1·0
–0·7
–0·5
–0·3
–0·1
0·1
0·3
0·5
0·7
Forest productivity
MEE species richness
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
(c) Centre vs. edge of clearing
0·0 0·2 0·4 0·6 0·8 1·0
−0·7
−0·5
−0·3
−0·1
0·1
0·3
0·5
0·7
Regrowth of trees in clearing
MEE species richness
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
(d) Centre vs. edge of clearing
Fig. 5. Influence of explanatory variables
on edge effects into (a) forest and (b), (c)
and (d) clearings. Increasing values of the
response in (a) and (b) mean larger differ-
ences in species composition between edge
and forest/clearing. Negative values for the
magnitude of edge effects (MEE) in (c)
and (d) mean that species richness was
lower in the edge than in the clearing cen-
tre.
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
878 K. Eldegard, Ø. Totland & S. R. Moe

we recorded in these centres have their primary habitat in
cultural landscapes (Eyitayo 2014). The conservation
value of early successional forest has recently received
increased attention (King & Schlossberg 2013). Power line
clearings can provide stable open-canopy habitats for
early successional bird and mammal species (Clarke &
White 2008; Askins, Folsom-O’Keefe & Hardy 2012) and
pollinating insects (Wojcik & Buchmann 2012). Thus,
proper management of vegetation in power line clearings
could positively affect diversity of other trophic levels.
One should not consider this system as binary –‘habitat’
(forest) and ‘non-habitat’ (clearing) –but rather as ‘habi-
tat’ and ‘alternative habitat’ (Haila 2002). Prioritizing
management practices that mitigate reductions in species
richness, diversity and important functional groups (e.g.
forbs) towards the edge, for instance by reducing the
amount of regrowth in the clearings, will promote the use
of clearings as replacement habitats. Negative edge effects
were also smaller at productive sites: ones that are proba-
bly of more interest as replacement habitats, because of
higher species richness and diversity, and more cultural
landscape species (Eyitayo 2014).
The difference in species composition between clearing
centres and clearing edges increased with power line age.
Immigration of non-forest species might cause temporal
edge effects on plant community composition; the value
of the clearings as replacement habitats could possibly
increase with power line age. Time-lags in the emergence
of fragmentation effects can confound effects in habitat
fragmentation studies (Ewers & Didham 2006b); lags of
50–100 years in plant diversity responses can occur (Lind-
borg & Eriksson 2004).
MANAGEMENT IMPLICATIONS
When planning new power lines, forested areas of high
conservation value should preferably be avoided. If new
power lines are to be established in such areas, the prior-
ity should be to minimize edge effects on the plant com-
munity across the edge and into the adjacent forest.
Specifically, one should avoid creating edges that face the
equator, bypass areas with high forest productivity,
choose landscapes that already have extensive forest edges
and minimize edge contrast. For existing power lines, we
recommend developing management plans that differenti-
ate between three types of power line clearings:
1. Clearings with adjacent forest of high conservation
value, where the priority should be to minimize edge
effects across the edge and into the adjacent forest. This
can be achieved by the following: (i) reducing the contrast
in height and density of the tree layer across the edge by
retaining more trees in a buffer zone along the edges of
clearings, with only selective manual felling; (ii) paying
particular attention to the above management practice on
edges facing the equator, at sites with high forest produc-
tivity and in landscapes with a low degree of forest
fragmentation; and (iii) continuing the current practice of
cutting every 5–10 years, given our results showing that
the amount of regrowth in the clearing centres does not
influence edge effects into forests.
2. Clearings that can act as replacement habitats, used to
increase local species richness and biodiversity, and which
can serve as alternative habitats for vanishing cultural
landscape species. Maintaining open-canopy habitats
under power lines can also improve conditions for other
groups of organisms, such as pollinating insects. Specifi-
cally, we suggest maintaining open-canopy habitats
through frequent cutting, to retain high richness and
diversity in the whole breadth of the corridor, thereby
mitigating negative edge effects. Such actions are particu-
larly important on highly productive sites because these
are richer, more diverse and have more cultural landscape
species, which could become endangered if cultural land-
scapes shrink.
3. ‘Business-as-usual’ clearings, where we recommend
continued maintenance cutting of all trees every 5–
10 years if there are no detrimental edge effects into
adjacent forest and the power line clearing is unsuitable
as replacement habitat for some plant or animal spe-
cies.
Acknowledgements
We thank the field assistants A. Kammerhofer, T. Kornstad, L. Nordtiller,
D. Slettebø, A.E. Rognes, E.S. Meen, M. Meland and A. Rasmussen; A.-
L. Aase for artwork in Fig. 1; V.V. Bischof and P.G.H. Frost for proofing
the English; Statnett for funding the field work; and J. Barlow, D. Hooft-
man and three anonymous referees for insightful and constructive com-
ments on previous drafts of the manuscript. The authors declare that they
have no conflict of interest.
Data accessibility
Plant community and environmental data available from the Dryad Digi-
tal Repository http://dx.doi.org/10.5061/dryad.5gg87 (Eldegard, Totland &
Moe 2015).
References
Askins, R.A., Folsom-O’Keefe, C.M. & Hardy, M.C. (2012) Effects of
vegetation, corridor width and regional land use on early successional
birds in power-line corridors. PLoS ONE,7, e31520.
Baker, S.C., Spies, T.A., Wardlaw, T.J., Balmer, J., Franklin, J.F., &
Jordan, G.J. (2013) The harvested side of edges: effect of retained for-
ests on the re-establishment of biodiversity in adjacent harvested areas.
Forest Ecology and Management,302, 107–121.
Barnes, A.D., Emberson, R.M., Chapman, H.M., Krell, F.-T. & Didham,
R.K. (2014) Matrix habitat restoration alters dung beetle species
responses across tropical forest edges. Biological Conservation,170,28–
37.
Belyea, L.R. & Lancaster, J. (1999) Assembly rules within a contingent
ecology. Oikos,86, 402–416.
Bengtsson, J., Nilsson, S.G., Franc, A. & Menozzi, P. (2000) Biodiversity,
disturbances, ecosystem function and management of European forests.
Forest Ecology and Management,132,39–50.
Berg,

A., Ahrn
e, K.,
€
Ockinger, E., Svensson, R. & S€
oderstr€
om, B. (2011)
Butterfly distribution and abundance is affected by variation in the
Swedish forest-farmland landscape. Biological Conservation,144, 2819–
2831.
Braithwaite, N.T. & Mallik, A.U. (2012) Edge effects of wildfire and ripar-
ian buffers along boreal forest streams. Journal of Applied Ecology,49,
192–201.
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
Edge effects along power line clearings 879

Burt, J.W. & Rice, K.J. (2009) Not all ski slopes are created equal: distur-
bance intensity affects ecosystem properties. Ecological Applications,19,
2242–2253.
Chabrerie, O., Jamoneau, A., Gallet-Moron, E. & Decocq, G. (2013)
Maturation of forest edges is constrained by neighbouring agricultural
land management. Journal of Vegetation Science,24,58–69.
Chao, A., Jost, L., Chiang, S.C., Jiang, Y.-H. & Chazdon, R.L. (2008) A
two-stage probabilistic approach to multiple-community similarity indi-
ces. Biometrics,64, 1178–1186.
Clarke, D.J. & White, J.G. (2008) Towards ecological management of
Australian powerline corridor vegetation. Landscape and Urban Plan-
ning,86, 257–266.
Driscoll, D.A., Banks, S.C., Barton, P.S., Lindenmayer, D.B. & Smith,
A.L. (2013) Conceptual domain of the matrix in fragmented landscapes.
Trends in Ecology and Evolution,28, 605–613.
Dumas, B.C., Holmstead, G.L., Kerr, M.J.J. & Carpenter, L.B. (2003)
Effects of road and transmission-line rights-of-way on botanical resources.
Idaho Power Company, Technical Report Appendix E.3.3-4. Idaho
Power Company, Id.
Eldegard, K., Totland, Ø. & Moe, S.R. (2015) Data from: edge effects on
plant communities along power line clearings. Dryad Digital Repository,
http://dx.doi.org/10.5061/dryad.5gg87
ESRI (2011) ArcGIS Desktop: Release 10. Environmental Systems
Research Institute, Redlands, CA.
Ewers, R.M. & Didham, R.K. (2006a) Continuous response functions for
quantifying the strength of edge effects. Journal of Applied Ecology,43,
527–536.
Ewers, R.M. & Didham, R.K. (2006b) Confounding factors in the detec-
tion of species responses to habitat fragmentation. Biological Review,
81, 117–142.
Eyitayo, D.L. (2014) Shifts in plant communities caused by vegetation clear-
ing in power-line corridors. MSc thesis, Norwegian University of Life
Sciences.
Fahrig, L. (2003) Effects of habitat fragmentation on biodiversity. Annual
Review of Ecology and Systematics,34, 487–515.
Fonseca, C.R. & Joner, F. (2007) Two-sided edge effect studies and the
restoration of endangered ecosystems. Restoration Ecology,15, 613–619.
Forman, R.T.T. (1995) Land Mosaics. The Ecology of Landscapes and
Regions. Cambridge University Press, Cambridge, UK.
Forman, R.T.T., Sperling, D., Bissonette, J.A., Clevenger, A.P., Cutshall,
C.D., Dale, V.H. (2003) Road Ecology. Science and Solutions. Island
Press, Washington, DC.
Forrester, J.A., Leopold, D.J. & Hafner, S.D. (2005) Maintaining critical
habitat in a heavily managed landscape: effects of power line corridor
management on Karner blue butterfly (Lycaeides melissa samuelis)
habitat. Restoration Ecology,13, 488–498.
Haila, Y. (2002) A conceptual genealogy of fragmentation research: from island
biogeography to landscape ecology. Ecological Applications,12, 321–334.
Hamberg, L., Lehvavirta, S. & Kotze, D.J. (2009) Forest edge structure as
a shaping factor of understorey vegetation in urban forests in Finland.
Forest Ecology and Management,257, 712–722.
Harper, K.A., Macdonald, S.E., Burton, P.J., Chen, J., Brosofske, K.D.,
Saunders, S.C. et al. (2005) Edge influence on forest structure and com-
position in fragmented landscapes. Conservation Biology,19, 768–782.
Hill, M.O., Mountford, J.O., Roy, D.B. & Bunce, R.G.H. (1999) ECO-
FACT 2a Technical Annex –Ellenberg’s Indicator Values for British
Plants. ITE, Natural Environment Research Council, Huntington, UK.
Hylander, K. (2005) Aspect modifies the magnitude of edge effects on bryo-
phyte growth in boreal forests. Journal of Applied Ecology,42, 518–525.
Jules, E.S. & Shahani, P. (2003) A broader ecological context to habitat
fragmentation: why matrix habitat is more important than we thought.
Journal of Vegetation Science,14, 459–464.
King, D.I. & Schlossberg, S. (2013) Synthesis of the conservation value of
the early-successional stage in forests of eastern North America. Forest
Ecology and Management,324, 186–195.
Laurance, W.F., Goosem, M. & Laurance, S.G.W. (2009) Impacts of
roads and linear clearings on tropical forests. Trends in Ecology and
Evolution,24, 659–669.
Laurance, W.F., Nascimento, H.E.M., Laurance, S.G., Andrade, A.,
Ewers, R.M., Harms, K.E., Luiz~
ao, R.C.C. & Ribeiro, J.E. (2007) Hab-
itat fragmentation, variable edge effects, and the landscape-divergence
hypothesis. PLoS ONE,2, e1017.
Lindborg, R. & Eriksson, O. (2004) Historical landscape connectivity
affects present plant species diversity. Ecology,85, 1840–1845.
Lindenmayer, D.B. & Fischer, J. (2006) Habitat Fragmentation and Land-
scape Change. An Ecological and Conservation Synthesis. Island Press,
Washington, DC.
Lindenmayer, D.B. & Franklin, J.F. (2002) Conserving Forest Biodiversity.
A Comprehensive Multiscaled Approach. Island Press, Washington, DC.
Luken, J.O., Hinton, A.C. & Baker, D.G. (1991) Forest edges associated
with power-line corridors and implications for corridor siting. Land-
scape and Urban Planning,20, 315–324.
Luken, J., Hinton, A.C. & Baker, D.G. (1992) Response of woody plant-
communities in power-line corridors to frequent anthropogenic distur-
bance. Ecological Applications,2, 256–362.
McWethy, D.B., Hansen, A.J. & Verschuyl, J.P. (2009) Edge effects for
songbirds vary with forest productivity. Forest Ecology and Manage-
ment,257, 665–678.
Nekola, J.C. (2012) The impact of a utility corridor on terrestrial gastro-
pod biodiversity. Biodiversity Conservation,21, 781–795.
Nilsen, P. & Larsson, J.Y. (1992) Bonitering av skog ved hjelp av vegetasj-
onstype og egenskaper ved voksestedet. (Site Index Estimation from
Vegetation Type and Site Properties.) Rapport fra Skogforsk, 22/92,
pp. 1–43, Norsk Institutt for Skogforskning,

As, Norway (in Norwegian
with English summary).
Økland, R.H., Økland, T. & Rydgren, K. (2001) Vegetation-environment
relationships of boreal spruce swamp forests in Østmarka Nature
Reserve, SE Norway. Sommerfeltia,29,1–190.
Pellissier, V., Berg
es, L., Nedeltcheva, T., Schmitt, M.-C., Avon, C.,
Cluzeau, C. & Dupouey, J.-L. (2013) Understorey plant species show
long-range spatial patterns in forest patches according to distance-
to-edge. Journal of Vegetation Science,24,9–24.
Pickett, S.T.A., Cadenasso, M.L. & Meiners, S.J. (2013) Vegetation
dynamics. Vegetation Ecology, 2nd edn (eds E. van der Maarel & J.
Franklin), pp. 107–140. John Wiley & Sons, Chichester, UK.
Pinheiro, J., Bates, D., DebRoy, S. & Sarkar, D. & R Core Team (2014)
nlme: Linear and Nonlinear Mixed Effects Models. R package version
3.1-117, http://CRAN.R-project.org/package=nlme.
Pohlman, C.L., Turton, S.M. & Goosem, M. (2009) Temporal variation in
microclimatic edge effects near power-lines, highways and streams in
Australian tropical rainforests. Agricultural and Forest Meteorology,
149,84–95.
Powell, A.S. & Lindquist, E.S. (2011) Effects of power-line maintenance
on forest structure in a fragmented urban forest, Raleigh, NC. South-
eastern Naturalist,10,25–38.
Russell, K.N., Ikerd, H. & Droege, S. (2005) The potential conservation
value of unmowed powerline strips for native bees. Biological Conserva-
tion,124, 133–148.
SAS (2011) SAS/STAT
â
9.3 User’s Guide: The GLIMMIX Procedure
(Chapter). SAS Institute Inc., Cary, NC.
Wojcik, V.A. & Buchmann, S. (2012) Pollinator conservation and manage-
ment on electrical transmission and roadside rights-of-way: a review.
Journal of Pollination Ecology,7,16–26.
Received 15 December 2014; accepted 29 April 2015
Handling Editor: Danny Hooftman
Supporting Information
Additional Supporting Information may be found in the online version
of this article.
Fig. S1. Details on GNMDS ordination of the plant data.
Table S1. Observed means for response variables in forest interior
plots.
Table S2. Variables influencing edge effects on either side of the
edge.
©2015 The Authors. Journal of Applied Ecology ©2015 British Ecological Society, Journal of Applied Ecology,52, 871–880
880 K. Eldegard, Ø. Totland & S. R. Moe