ArticlePDF Available

A Comparison of All-Terrain Vehicle (ATV) Trail Impacts on Boreal Habitats Across Scales


Abstract and Figures

Recreational trails are an agent of anthropogenic disturbance in nature reserves and other low human impact areas. Effective management must balance the desire of recreationists to use these natural areas with the need to maintain their ecological integrity. Environments with low productivity may be particularly susceptible due to low resistance and resilience to recreational impacts. Our study examined 28 all-terrain vehicle (ATV) trails within the Avalon Wilderness Reserve and the adjacent surrounding area on the island of Newfoundland, Canada. We found that different habitat types (boreal forest, heaths, and bogs) differ in resistance and resilience to both direct on-trail erosion and indirect off-trail vegetation impacts of ATV trails. Dry forested sites were more resistant to direct on-trail erosion but less resistant to indirect off-trail vegetation disturbance. Heath sites were less resistant to direct on-trail erosion but highly resistant to indirect off-trail disturbance. Bog sites had low resistance to both direct and indirect trail disturbance. There have been limited studies on ATV trail impacts in boreal environments, and our findings provide guidance for managers in such environments to manage recreational vehicle use.
NMDS ordinations of life form community level analysis among factors. Triangles denote quadrats, numbering indicates their spatial position along the line transect. The number 1 denotes the closest position to the trail (i.e., the edge) and the number 11 denotes the further position from the trail (i.e., the interior). Crosses denote group centroids. Triangles that are close together have more similar life form assemblages than triangles that are further apart. a. Comparison of forest samples (closed triangles ▲) to open samples (open triangles ∆). The 2-dimensional solution explained 97.0% of the variation, with a final stress of 6.196, achieved after 53 iterations (Monte Carlo stress test P = 0.004). b. Comparison of dry samples (closed triangles ▲) to wet samples (open triangles ∆). The 2-dimensional solution explained 92.0% of the variation, with a final stress of 10.336, achieved after 45 iterations (Monte Carlo stress test P = 0.004). c. Comparison of non-reserve samples (closed triangles ▲) to reserve samples (open triangles ∆). The 2-dimensional solution shown explains 88.7% of the variation, however NMDS ordination yielded a 3-dimensional solution explaining 96.9% of the variation with a final stress of 5.359 after 70 iterations (Monte Carlo stress test P = 0.004). d. Comparison of forested dry samples (closed triangles▲), forested wet samples (open triangles ∆), open dry samples (inverted open triangles ∇), and open wet samples (inverted closed triangles ▼). The 2-dimensional solution shown explains 78.9% for the variation, however, NMDS yielded a 3-dimensional solution explaining 94.5% of the variation with a final stress of 8.211 after 69 iterations (Monte Carlo stress test P = 0.004).
Content may be subject to copyright.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research
libraries, and research funders in the common goal of maximizing access to critical research.
A Comparison of All-Terrain Vehicle (ATV) Trail Impacts on Boreal Habitats
Across Scales
Author(s): Nyssa van Vierssen Trip and Yolanda F. Wiersma
Source: Natural Areas Journal, 35(2):266-278.
Published By: Natural Areas Association
BioOne ( is a nonprofit, online aggregation of core research in the biological, ecological, and
environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published
by nonprofit societies, associations, museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of
BioOne’s Terms of Use, available at
Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries
or rights and permissions requests should be directed to the individual publisher as copyright holder.
266  Natural Areas Journal  Volume 35 (2), 2015
ABSTRACT: Recreational trails are an agent of anthropogenic disturbance in nature reserves and other
low human impact areas. Effective management must balance the desire of recreationists to use these
natural areas with the need to maintain their ecological integrity. Environments with low productivity
may be particularly susceptible due to low resistance and resilience to recreational impacts. Our study
examined 28 all-terrain vehicle (ATV) trails within the Avalon Wilderness Reserve and the adjacent
surrounding area on the island of Newfoundland, Canada. We found that different habitat types (boreal
forest, heaths, and bogs) differ in resistance and resilience to both direct on-trail erosion and indirect
off-trail vegetation impacts of ATV trails. Dry forested sites were more resistant to direct on-trail
erosion but less resistant to indirect off-trail vegetation disturbance. Heath sites were less resistant to
direct on-trail erosion but highly resistant to indirect off-trail disturbance. Bog sites had low resistance
to both direct and indirect trail disturbance. There have been limited studies on ATV trail impacts in
boreal environments, and our findings provide guidance for managers in such environments to manage
recreational vehicle use.
Index terms: all-terrain vehicle, boreal habitats, edge effects, erosion, recreation ecology
In general, environments with low produc-
tivity, such as boreal habitats, have low
resilience and resistance to recreational
vehicle impacts (Allard 2003). Here, we
adopt the definitions by Kuss and Hall
who characterized plant resistance to
recreational impacts as the ability to with-
stand being “injured or impaired” (Kuss
and Hall 1991), while resilience was the
ability of a plant that had been trampled
to “survive or regenerate” (Kuss and Hall
1991). In addition to primary productivity,
plant morphology is another important
determinant of resistance and resilience.
Plants possessing traits such as durable and
flexible stems and leaves, or small (close to
the ground) compact growth form, or below
ground reproductive structures generally
have higher resistance (Cole 1995; Liddle
1997; Yorks et al. 1997). However, environ-
mental productivity and plant morphology
may result in confounding responses. Soil
moisture compounds these impacts; gener-
ally, the greater the soil moisture the greater
the impact. The overall vegetative cover
also plays a role, with denser canopies
having understory plants that are more
susceptible to recreation impacts. Adapta-
tions of shade-tolerant understory plants
(i.e., greater leaf area, thinner cuticles)
make them less resistant to recreational
trampling (Cole 1978, 1995). Ground layer
properties also play a role, and conifer-
ous forests may be less vulnerable than
deciduous ones due to their relatively large
unincorporated organic litter (Legg 1973
as cited in Kuss 1986).
Within the field of recreation ecology it has
been recognized that spatial scale aspects of
motorized recreational impacts have been
largely understudied (Cole 2004; Brooks
and Lair 2005; Ouren et al. 2007). Brooks
and Lair (2005) define three distinct spatial
scales at which vehicular trail impacts oc-
cur: (1) direct effects; (2) indirect effects;
and (3) landscape effects. Direct effects
occur within the trail itself (Brooks and
Lair 2005), such as loss of vegetation cover
or erosion through rutting. The majority of
research on ORV (Off Road Vehicle) trails
has been direct effect based (Liddle 1997;
Leung and Marion 2000; Buckley 2004;
Ouren et al. 2007). Indirect effects occur
in areas adjacent to trails (Brooks and
Lair 2005) such as increased sediment or
nutrient loading of surrounding vegetation.
As the name implies, landscape effects
are dispersed throughout the landscape
(Brooks and Lair 2005) (e.g., habitat
fragmentation, spread of invasive species)
and can be difficult to quantify. Indirect
and landscape effects are generally very
context-specific as they are influenced
by specific environmental and ecological
gradients and different land use regimes
(Brooks and Lair 2005). Nevertheless,
elucidating ORV impacts at the appropri-
ate spatial scale of ecosystem response is
crucial for efficient management of these
impacts (Brooks and Lair 2005), particu-
larly when many management decisions
are at the landscape level.
Understanding ORV impacts to assist
with management is important even in
places where levels of vehicular traffic
are low, since trails can be created and
Natural Areas Journal 35:266–278
2 Corresponding author:; 709-864-7499
A Comparison of
All-Terrain Vehicle
(ATV) Trail Impacts
on Boreal Habitats
Across Scales
Nyssa van Vierssen Trip1
1Department of Biology
Memorial University
St. John’s, Newfoundland and Labrador,
Canada, A1B 3X9
Yolanda F. Wiersma1,2
Volume 35 (2), 2015 Natural Areas Journal 267
have substantial ecological impact even
where there is light traffic (Weaver and
Dale 1978). This is due to the curvilinear
use-impact relationship (Liddle 1997; Cole
2004; Quinn and Chernoff 2010), which
has demonstrated that impacts can be found
at low intensities of use and then reach a
threshold at high intensities of use.
Here, we present a study in which we
examined direct effects (on-trail impact)
measured by soil erosion and indirect
effects (off-trail impact) measured by
changes in the vegetation community away
from trails in and around a boreal protected
area, the Avalon Wilderness Reserve. We
use the direct effect of erosion damage
as an indicator of system resilience, and
indirect vegetation impacts (edge effects)
as an indicator of resistance to impacts
from All-Terrain Vehicles (ATVs), a type
of ORV. As a protected area where ORV
(including ATV) use is prohibited, the
Avalon Wilderness Reserve was expected
to have lower ATV traffic volume on
existing trails than the adjacent area. Our
intent was to have the Wilderness Reserve
act as a contrast to the higher trafficked
nonprotected area, but where habitat types
were similar.
We hypothesized that ecological impacts
of ATVs as measured by on-trail erosion,
will be influenced by cover type (habitat),
moisture level (microhabitat), and intensity
of use. Low productivity habitats are be-
lieved to be highly susceptible to, and slow
to recover from, recreation disturbance
(Liddle 1975). Open low productivity
habitats such as deserts, tundra, and alpine,
have been shown to be highly susceptible
to erosion via recreational trails (Liddle
1997; Buckley 2004). Previous studies
demonstrated that hydric and mesic sites
are more susceptible to erosion than more
xeric sites (Radforth 1972; Burton 1974 as
cited in Liddle 1975; Jones 1978 as cited
in Yorks et al. 1997; Lagocki 1978 as
cited in Liddle 1997). Thus, we predicted
that: (1a) on-trail erosion will be greater
in open habitats than in forested habitats;
(1b) on-trail erosion will be greater in
wet sites than dry sites; and (1c) on-trail
erosion will be greater on high trafficked
trails than on low trafficked trails.
Recreation impacts on vegetation both on-
and off-trail include crushing, abrasion,
introduction of exotic/invasive species,
overall reduction of biomass (particularly
of sensitive species), and shifts in species
composition (Liddle 1997; Cole 2004;
Rooney 2005; Pickering and Hill 2007).
Shifts in species composition may include
a shift to more nonnative species, and/or to
a community dominated by those species
that can withstand trampling and physi-
cal disturbance (Liddle 1997; Cole 2004;
Rooney 2005; Pickering and Hill 2007).
Our second hypothesis was that ecological
impacts of ATVs as measured by off-trail
impacts (edge effects) on vegetation will be
influenced by cover type (habitat), moisture
level (microhabitat), and intensity of use.
Open habitats such as tundra, heath, and
bogs have low resilience and long recovery
times following recreational (particularly
vehicular) disturbance (Willard and Marr
1970; Greller et al. 1974; Bayfield 1979;
Charman and Pollard 1993). Wetter areas
within a given habitat type have shown
more erosion and are likely to be denuded
of vegetation more quickly than drier areas
(Willard and Marr 1970; Radforth 1972;
Liddle 1997; Törn et al. 2009; Jorgenson et
al. 2010) given similar types and amounts
of use. We predicted that: (2a) off-trail
vegetation impacts (changes in species
composition) will be greater (appear fur-
ther from the trail) in open habitats than in
forested habitats; (2b) off-trail vegetation
impacts (changes in species composition)
will be greater (appear further from the
trail) in wet sites than dry sites; and (2c)
off-trail vegetation impacts (changes in spe-
cies composition) will be greater (appear
further from the trail) on high trafficked
trails than low trafficked trails.
Study Area
Our study area was located in the Maritime
Barrens Ecoregion (MBE, Damman 1983)
on the east central section of the Avalon
Peninsula, on the island of Newfound-
land, Canada (47o6’ N, 53o15’ W) (Figure
1). The MBE falls within the southern
boreal ecozone, and is characterized by
cool summers (mean temperature 13 to16
oC) and mild winters (mean temperature
-3 to -8 oC) (Damman 1983), with high
precipitation (over 1250 mm annually)
and frequent fog cover (Damman 1983).
The MBE is dominated by nutrient poor
environments such as heaths, bogs, and
fens, with forest stands occurring in more
sheltered areas such as valleys (Damman
1983). Heath vegetation is dominated
by the dwarf shrub Kalmia augustifolia
L. (sheep laurel), however Rhododenron
canadense (L.) Torr. (rhodora) and Vac-
cinium augustifolim Ait. (wild blueberry)
are also abundant (Damman 1983). Other
dominant ground cover includes members
of the lichen genus Cladonia spp., mosses
Pleurozium schreberi (Brid.) Mitt. and
Sphagnum spp. (in wetter areas) (Dam-
man 1983; Meades 1983). Forests are
dominated by black spruce (Picea mariana
Mill.) and balsam fir (Abies balsamea L.)
(Damman 1983).
The study area included the Avalon Wil-
derness Reserve (AWR) and the adjacent
area. The AWR is 1070 km2 in size and
contains a single unpaved road (Figure 1)
that penetrates deeply into the reserve. The
road is privately owned and maintained
by Newfoundland Power Incorporated to
provide access to dams in the area, and
its construction predates the founding of
the reserve (Avalon Wilderness Reserve
Management Plan 1986). The road is
suitable for 4-wheel drive vehicles and
ATVs only. Although ATV use is officially
prohibited within the reserve, there is very
little enforcement, and local information
suggested that there was a high probability
of discovering illegal trails.
Study Design
Our study design incorporated a hierarchi-
cal 3x2 factorial design. The three factors
were: (1) legal status (a nonprotected
area, Non-Reserve, and a protected area,
Reserve; hereafter NR/R); (2) habitat
(Forest/Open; hereafter F/O); and (3) mi-
crohabitat (Dry/Wet; hereafter D/W). Open
environments include bogs and heaths. We
used 55 cm aerial ortho photographs taken
in summer 2008 to identify potential ATV
trails (data source: Newfoundland and Lab-
rador Department of Natural Resources).
268 Natural Areas Journal Volume 35 (2), 2015
Figure 1. Map of the general study area location on the southeast Avalon Peninsula, island of Newfoundland, Canada. a. The location of the province of New-
foundland and Labrador (shaded grey) within North America. b. The island of Newfoundland; the Avalon Peninsula is shaded in grey. c. Detail of the Avalon
Peninsula. The Avalon Wilderness Reserve (AWR) is crosshatched, with the study area outside the reserve (non-reserve; NR) in grey. Horse Chops road is the
black line and the only road into the AWR.
Volume 35 (2), 2015 Natural Areas Journal 269
Further trails were discovered during initial
site visits. We assigned treatments using a
random stratified sample in a Geographic
Information System (GIS; ArcGIS version
9.3) with data sources that represented
cover and moisture levels (see van Viers-
sen Trip 2014 for details). All treatment
assignments from the random stratification
were verified at the beginning of the field
season (May 2010); where necessary, new
sampling locations were chosen or factors
reassigned. We mapped all trails using a
handheld Geographic Positioning System
(GPS) GARMIN 76 and walked the entire
length of each trail. There were a total of
28 trails, with lengths ranging from 29 to
1415 m (mean 459 m).
data Collection
We deployed magnetic Off-Highway Ve-
hicle (OHV) counters (G3 OHV counters
manufactured by TRAFx Research Ltd.
2012) on a subset of trails. We rotated
counters throughout the study period to
obtain replicates for all factors. Counters
were deployed at 16 of 28 of the trails;
however one counter failed to start, so
data from only 15 trails were collected.
The average length of time a counter was
placed at any one trail was 19 (± 10 SD)
days. Because we could not deploy counters
on all trails, nor during the entire season,
we also noted what was at the end of each
trail (the “destination”) as a potential proxy
for traffic intensity. Destinations were lakes
in 15 cases (which may have higher traffic
to access fishing in early spring before we
deployed counters), domestic cutovers in
three cases (which may have zero to low
traffic if no longer in active use), campsites
in six cases, circling back to the road in two
cases, and overgrowth in two cases.
On longer trails (>300 m) we took erosion
depth measurements (rut depth measure-
ments) every 100 meters. On shorter trails
(<300 m) we took depth measurements
at the beginning, midpoint, and end of a
trail. We used a measuring tape to take
measurements, which were rounded to
the nearest millimeter, and calculated the
average amount of erosion for each trail
from all samples.
We established 50-m line transects begin-
ning at the midpoint of each trail, per-
pendicular to a trail. At every five meters
starting from the edge (beside wheel ruts
or a visible path), we laid 1-m2 quadrats.
Quadrats were placed with the lower left
corner at the appropriate meter demarca-
tion. Within each quadrat we identified all
vascular species and estimated their percent
cover. We categorized nonvascular species
broadly as either moss or lichen, recorded
their presence/absence, and estimated their
percent cover.
Statistical Analysis
Spatial Autocorrelation
Because sites were not chosen completely
at random, we applied join count statistics
to examine spatial autocorrelation among
trails based on experimental factors (NR/R;
F/O, D/W) (Fortin and Dale 2005). We
used GPS coordinates (point data) taken
at the beginning of each trail to act as a
location of each trail. Protected status was
not tested for spatial autocorrelation for
two reasons: (1) it was a legal, rather than
biological, factor and, therefore, we were
less concerned with possible confounding
effects; and (2) by definition, trails within
one of the two conditions (NR or R) will
be close together spatially and likely
positively autocorrelated. We used a binary
weighing matrix to define the strength of
connections between sample points (Cliff
and Ord 1981) because we did not have
a priori assumptions about the strength of
connections between sampling locations.
We used nearest neighbor connectivity
to define the degree of spatial proximity
between the sample locations. We ran
the join count statistic under non-free
sampling assumption (Sokal and Oden
1978). Statistical analysis was carried out
using R 2.14.1 Statistical Software (R Core
Team 2011).
We tested differences in ATV traffic level
among and between factors (legal status,
habitat, and microhabitat) with a non-
parametric Kruskal-Wallis ANOVA in R
(version 2.14.1; R Core Team 2011). We
also included trail destination in GzLMs
to test whether this had an effect.
To test the relationship between erosion
depth and the legal status, habitat, and
microhabitat, we analyzed the data using
Generalized Linear Models (GzLMs) in R
(version 2.14.1; R Core Team 2011). All
GzLMs were Gaussian distributed and used
the identity link. The response variable
(erosion depth) was normalized via log
transformation prior to analysis. Predictor
variables included the a priori factors as
well as the post hoc variable of destination
(i.e., trails that end in lakes compared to
all other types of endpoints). We compared
models using the information theoretic
approach with the corrected Akaike’s In-
formation Criterion (AICc) to account for
low ratio of sample size to parameters
(Burnham and Anderson 2002).
To test if our experimental groups (i.e.,
NR/R, F/O, D/W) were truly capturing
differences in vegetation community com-
position, we performed a Multi-Response
Permutation Procedure (MRPP) using PC-
ORD version 6.0 (McCune and Mefford
2011). MRPP is a nonparametric statisti-
cal technique that tests for a difference
among two or more groups in one or more
dimensions (Mielke and Berry 2007). We
used a natural weighting as recommended
by Mielke (1984) and a Sørensen (Bray-
Curtis) distance metric to measure the
difference in ecological distance between
factors. We also rank transformed the
distance matrix to accommodate the fact
that as ecological community heterogene-
ity increases, distance metrics can suffer a
loss of sensitivity (Clarke 1993; McCune
and Grace 2002). We tested for differences
within and among the three factors (legal
status, habitat, and microhabitat). We
270 Natural Areas Journal Volume 35 (2), 2015
pooled data by summing abundances across
quadrats at the same position along the line
transect within a given habitat (e.g., all
quadrats across trails in a forest site at the
5-m mark were pooled). Thus, when data
were pooled, relative spatial position was
maintained. Identical analyses described
below were also run on unpooled data.
To examine community gradients (i.e.,
indirect effects) we used Non-Metric
Multidimensional Scaling (NMDS) ordi-
nation and Polar (Bray-Curtis) ordination
using PC-ORD version 6.0 (McCune and
Mefford 2011). NMDS is a “free ordina-
tion” technique (Minchin 1987; Clarke
1993; Peck 2010). To ordinate sites, we
calculated a Sørenson (Bray-Curtis) dis-
similarity matrix on species and life form
abundance data (see van Vierssen Trip 2014
for complete list of species and life form
classification). Data were not transformed
prior to analysis; the NMDS algorithm
does not assume linearity among variables
(McCune and Grace 2002). We performed
NMDS stepping-down (beginning at the
highest dimension) 1 through 6 dimensions,
three separate times for each experimental
treatment comparison. Each step-down run
had a random starting configuration. We
set the iteration maximum to 250, and the
stability criterion <0.00001. We determined
appropriate dimensionality through the
inspection of scree plots, Monte Carlo tests
(250 iterations) of each dimensionality, and
inspecting the final stress of each dimen-
sion. Upon determining the appropriate
dimensionality, we performed NMDS five
times (with the above parameters); each
run had a random starting configuration.
Of the five runs, the starting coordinates
of lowest stress solution were used as the
starting coordinates for the final solution.
The final solution was inspected against
previous runs and, if no discrepancies were
found, was considered the global optima.
Plots had axes rotation making principal
axes orthogonal.
To test for the presence of the experimen-
tal gradient (i.e., indirect trail effect) we
performed a polar (Bray-Curtis) ordination.
In contrast to NMDS, polar ordination is a
“guided ordination” technique that assumes
the presence of an ecological community
gradient (Beals 1984; Peck 2010). We used
the Sørenson (Bray-Curtis) distance index.
We selected the closest sample to the trail
(i.e., the physical edge) as the first refer-
ence point and the furthest sample from
the trail (i.e., the interior) as the second
reference point.
Spatial Autocorrelation
Among the experimental factors, Forest/
Forest had the highest amount of positive
spatial autocorrelation. Forest-Dry/Forest-
Dry and Wet/Wet also had a notable level
of positive autocorrelation. In general, there
was not a high degree of positive autocor-
relation among sites of the same treatment
factor. Therefore, we are confident that our
sampling locations are reasonably spatially
Overall erosion depth ranged from 3.25
to 21.5 cm, mean = 11.21 cm, SD = 4.83
cm. The amount of erosion differed sig-
nificantly among factors. Forested trails
had significantly less erosion than open
trails (t(26, 27) = 20.394, P < 0.001; Fig-
ure 2a). Dry trails had significantly less
erosion than wet trails (t(26, 27) = 21.077,
P < 0.001; Figure 2b). Thus, both hy-
potheses 1a and 1b, that on-trail erosion
was predicted to be higher on open and
wetter trails respectively, were supported.
With the exception of dry-forested trails,
the other combinations of microhabitat
and habitat did not differ significantly in
erosion amount from one another (wet
forested trails (t(24, 27) = 0.435, P = 0.667,
dry open trails t(24, 27) = 0.385, P = 0.704,
and wet open trails t(24, 27) = 1.129, P =
0.270; Figure 2c); dry-forested trails had
significantly less erosion compared to all
other habitat types (t(24, 27) = 16.931, P <
0.001; Figure 2c). Trails within the AWR
showed a significant difference in erosion
level compared to trails outside the AWR
(t(26, 27) = 20.332, P < 0.001) but this dif-
ference became highly nonsignificant when
habitat type was accounted for (t(23, 27) =
0.301, P = 0.766).
Models that included microhabitat only
and habitat only ranked as top models in
model selection (i.e., had Δi < 2; Table 1).
The microhabitat only model had slightly
more weight of evidence (wi = 0.4357)
compared to the habitat only model (wi =
0.3985) (Table 1).
All trails had vehicular traffic. Traffic
counts per day ranged from 0.2 to 13.6
(mean = 3.18 and SD = 3.62). There was
no difference in the amount of traffic by
habitat (χ2 = 0.8602, df = 1, P = 0.3537),
by microhabitat (χ2 = 0.0538, df = 1, P =
0.8166), by legal status (χ2 = 0.1647, df
= 1, P = 0.6849) or among the different
habitat/microhabitat combinations (χ2 =
1.4692, df = 3, P = 0.6894). Because there
were no statistically significant differences
in traffic volume, we could not test whether
traffic volume affected erosion (hypothesis
1c) or edge effects (hypothesis 2c). When
using destination as a proxy for use, we
found only one difference; trails that ended
in lakes had higher erosion levels compared
to trails that ended in other destinations
(i.e., camp sites, wood cutting areas), t(26,
27) = 21.606, P < 0.001. However, this
result became highly nonsignificant when
habitat type was taken into account (t(23,
27) = 1.377, P = 0.182).
MRPP analysis confirmed that all pooled
experimental groups were significantly
different from one another both in com-
munity species composition and life form
composition (Table 2). Separation trends
were similar at both the species and life
form level of analysis, with most separa-
tion between groups among the habitat
types (T = -16.200, P < 0.001 [species],
T = -17.655, P < 0.001 [life form]) and
least separation between reserve and non-
reserve (T = -7.413, P < 0.001 [species], T
= -7.221, P < 0.001 [life form]). Greatest
dissimilarity in species composition among
the habitat types confirmed the assumption
of community level differences between the
experimental factors. Least dissimilarity
in species composition between the area
within and outside the reserve confirms
the assumption about this experimental
Volume 35 (2), 2015 Natural Areas Journal 271
factor that these two areas (levels) should
differ in legal status only. Chance-corrected
within group agreement, A, did not differ
markedly between the species and life
form level analyses (Table 2). Factors had
similar levels of community homogeneity
at species and life form level analyses
(Table 2). Similar results were seen for
the MRPP analysis using the unpooled
data (van Vierssen Trip 2014).
NMDS ordination of abundance by life
form showed a distinct separation between
habitat types (Figure 3a), indicating distinct
vegetation community types. Samples
from open sites were more clustered than
samples from forested sites (Figure 3a), in-
dicating a more homogeneous community
in open habitats along the experimental
gradient. There was also a clear separation
between edges samples (i.e., forest quadrats
1, 2 and 3) and interior samples (i.e., forest
quadrats 9, 10, 11). For the open habitat,
edge samples were more tightly clustered
with mid-way samples, but there is a
clear separation of interior samples (open
quadrats 9, 10, 11). This indicates a slower
rate of species turnover with distance from
trail in open habitats compared to forested
habitats. NMDS ordination showed a dis-
tinct separation between moisture levels
(Figure 3b), indicating distinct vegetation
community types. Wet samples were more
clustered than dry samples (Figure 3b), in-
dicating a more homogeneous community
in wet sites along the experimental gradi-
ent; however, there was clear separation
of interior samples (wet quadrats 10, 11).
In dry sites, there was a clear separation
between edge samples (dry quadrats 1, 2,
3) and interior samples (dry quadrats 9,
10, 11), indicating a slower rate of spe-
cies turnover (with distance from trail) in
wet sites compared to dry sites (Figure
3c). Separation between reserve and non-
reserve samples was clear (Figure 3c).
There was some separation between AWR
edge samples (quadrats 1, 2, 3) and interior
samples (quadrats 9, 10, 11), but only inte-
rior samples (quadrats 8–11) segregated in
the nonprotected samples (Figure 3c). Both
the R and NR showed detectable species
turnover (with distance from the trail);
however, turnover was more pronounced in
the PA. The NMDS ordination of habitat/
microhabitat combinations had somewhat
less clear trends. Upon comparison of all
habitat types using NMDS ordination,
only Forest/Dry and Open/Wet showed
clear community gradients (Figure 3d).
NMDS ordination results do not support
hypotheses 2a and 2b since strongest edge
(species turnover gradient) was detected in
forested and dry samples. Similar results
were seen for NMDS analysis on species
abundance as for the results on life form
abundance shown here (van Vierssen Trip
2014). Polar ordination showed strong de-
tectable life form community gradients for
edge to interior across all factors (Figure
4). Across the various habitat types, life
form groups that showed the strongest
association with the edge were shrubs,
graminoids, and mosses. Life form groups
that had the weakest association with the
edge were ferns and lichens (Table 3).
Figure 2. Box and whisker plots of erosion depth (cm) for the different factors. Horizontal bar is the
median, box is the interquartile range, whiskers are the highest and lowest extremes, circles are outliers
(1.5–3 box lengths from either end). Asterisks denote significance at P < 0.001. a. Factor is habitat with
levels F (forest; n = 16) and O (open; n = 12). b. Factor is microhabitat with levels D (dry; n = 17), and
wet (W; n = 11). c. Habitat/microhabitat factors with levels F/D (forest dry, n = 12), forest wet (F/W, n
= 4), open/ dry (i.e., heath) (O/D, n = 5) and open/wet (i.e., bog) (O/W, n = 7).
272 Natural Areas Journal Volume 35 (2), 2015
The comparative aspect of this study
provides valuable information about recre-
ational trail impacts in boreal habitats that
have not been well studied previously. Our
results illustrate that within the Maritime
Barrens Ecoregion (MBE), high and low
productive habitats respond differently to
direct and indirect impacts of ATV trails
with respect to resistance and resilience.
Unexpectedly, protected status did not
influence the amount of vehicle traffic,
as trails within and outside the Wilder-
ness Reserve had statistically equivalent
amounts of use. More productive habitats
(forests) and drier areas were more resis-
tant to direct impacts of rut formation, but
had weak resistance to indirect impacts, as
reflected by stronger gradients in vegeta-
tion composition. Less productive habitats
(open sites) and wet areas showed lower
resistance to direct and indirect impacts.
Bogs were highly sensitive to both direct
and indirect impacts. Interestingly, heath
sites were more vulnerable to direct impacts
than dry forest, but were more resistant to
indirect impacts.
Intensity of Use
We did not detect significant differences in
traffic volume between and among trails
within the various habitat types or between
the trails in areas under different legal
protection. While unexpected, given that
ORV use is prohibited within the AWR,
and unrestricted outside the reserve, these
findings do serendipitously control the
predictor variable of intensity of use (i.e.,
traffic volume). This is an intriguing finding
since several studies have demonstrated
an influence of intensity of recreational
activity on on-trail (i.e., amount of erosion)
impacts particularly at low levels of use
(Weaver and Dale 1978; Iverson et al. 1981;
Meadows et al. 2008). However, due to the
curvilinear-use impact relationship, several
studies have found on-trail soil erosion to
be a poor predictor of level of use (Dale
and Weaver 1974; Cole 1992; Olive and
Marion 2009), further illustrating the value
of using traffic counters to understand the
potential influence of intensity of use.
Model Log-Likelihood K AIC
Microhabitat 7.5161 3 -8.0321 0 0.4357
Habitat 7.4268 3 -7.8535 0.1786 0.3985
Microhabitat + Habitat Main Effects 7.7165 4 -5.6938 2.3383 0.1353
Global model 7.7221 5 -2.717 5.3151 0.0305
Table 1. Model selection results for models that predict log-transformed erosion levels (cm) for ATV trails (n = 28) found within the Avalon Wilderness
Reserve and outside the reserve in the adjacent surrounding area. K indicates number of parameters and AICc is corrected Akaike Information Criteria
to account for low ratio of n:K. i values indicate the difference between that model and the model with the lowest AICc. Models with i < 2 are consid-
ered plausible. Weight of evidence (0–1) is indicated by wi. Data based on field work from June–September 2011 in and around the Avalon Wilderness
Reserve, island of Newfoundland, Canada.
T A p T A p
Habitat Type Forest 11
Open 11
Microhabitat Dry 11
Wet 11
Reserve Non-Reserve: 11
Reserve 11
Habitat/Microhabitat Dry Forest 11
Wet Forest 11 -16.200 0.482 <0.001 -17.654 0.480 <0.001
Dry Open 11
Wet Open 11
Species Life Form
-11.456 0.375 <0.001 -10.439 0.317 <0.001
-9.351 0.311 <0.001 -9.003 0.274 <0.001
-7.413 0.241 <0.001 -7.221 0.219
Table 2. Multi-Response Permutation Procedure comparisons of species and life form community composition among and between factors. Sample size for
each level is indicated by n. The T statistic is a measure of groups separation, the more negative the value, the greater the separation. A is the chance-cor-
rected within group agreement, a measure of group homogeneity. If A = 1 all items within the group are identical. If A = 0, heterogeneity within groups is
equal to chance. Data based on field work from June–September 2011 in and around the Avalon Wilderness Reserve, island of Newfoundland, Canada.
Volume 35 (2), 2015 Natural Areas Journal 273
On-Trail Impacts
Comparisons of levels of erosion among
forested vs. open and dry vs. wet trails
confirmed predictions that: (1a) erosion
would be higher in open habitats; and
(1b) erosion would be higher at wetter
sites. Rut depth was significantly deeper
on open trails compared to forested trails.
We could not directly compare this result
to previous literature since we are unaware
of comparable studies that compared soil
erosion in a boreal forest and heath (bar-
rens) and bog environments, however our
findings are consistent with other studies
that demonstrated that wetter sites gener-
ally have higher levels of erosion than drier
sites (Bellamy et al. 1971; Bliss and Wein
1972; Bryan 1977; Weaver and Dale 1978;
Wilson and Seney 1994; Kelleway 2005).
Drier soil has greater capacity to bear a
moving load (Marshall and Holmes 1979
as cited in Kuss 1986). However, when
examining habitat/microhabitat combina-
tions, we found that there was no significant
difference in erosion level between dry
open trails and wet open or forest trails.
This was an unexpected result. A possible
explanation for our result is soil type. The
soils in our study area are characterized
by coarse podzolic soils with thin till and
exposed bedrock (Roberts 1983). Thus,
these soils lack a thick organic layer and
Figure 3. NMDS ordinations of life form community level analysis among factors. Triangles denote quadrats, numbering indicates their spatial position along
the line transect. The number 1 denotes the closest position to the trail (i.e., the edge) and the number 11 denotes the further position from the trail (i.e., the
interior). Crosses denote group centroids. Triangles that are close together have more similar life form assemblages than triangles that are further apart. a.
Comparison of forest samples (closed triangles ) to open samples (open triangles ). The 2-dimensional solution explained 97.0% of the variation, with a
final stress of 6.196, achieved after 53 iterations (Monte Carlo stress test P = 0.004). b. Comparison of dry samples (closed triangles ) to wet samples (open
triangles ). The 2-dimensional solution explained 92.0% of the variation, with a final stress of 10.336, achieved after 45 iterations (Monte Carlo stress test P
= 0.004). c. Comparison of non-reserve samples (closed triangles ) to reserve samples (open triangles ). The 2-dimensional solution shown explains 88.7%
of the variation, however NMDS ordination yielded a 3-dimensional solution explaining 96.9% of the variation with a final stress of 5.359 after 70 iterations
(Monte Carlo stress test P = 0.004). d. Comparison of forested dry samples (closed triangles▲), forested wet samples (open triangles ), open dry samples
(inverted open triangles ), and open wet samples (inverted closed triangles ). The 2-dimensional solution shown explains 78.9% for the variation, however,
NMDS yielded a 3-dimensional solution explaining 94.5% of the variation with a final stress of 8.211 after 69 iterations (Monte Carlo stress test P = 0.004).
274 Natural Areas Journal Volume 35 (2), 2015
Figure 4. Polar ordinations of life form community level analysis within a factor. Triangles denote quadrats, and numbering indicates their spatial position
along the line transect. The number 1 denotes the closest position to the trail (i.e., the edge) and the number 11 denotes the further position from the trail
(i.e., the interior). Panels: a–h show the following levels for habitat and microhabitat factors, with percentage of variation explained in parentheses: a. Forest
(85.8%) b. Open (80.8%) c. Dry (81.5%) d. Wet (76.8%) e. Forest-Dry (79.7%) f. Forest-Wet (50.1%) g. Open-Dry (16.3%) h. Open-Wet (88.4%).
Volume 35 (2), 2015 Natural Areas Journal 275
are relatively erosion resistant, which may
account for our findings. Monz et al. (2013)
recognize that the curvilinear relationship
is not always applicable. Ecosystem re-
sponses to recreational use may be linear
or exponential; high resistant systems
may exhibit a flat (zero slope) relationship
with increasing use (Monz et al. 2013).
AIC analysis indicates that both habitat
and microhabitat were important drivers
of on-trail erosion, although microhabitat
had a slightly higher weight. This agrees
with the findings of previous researchers
(Weaver and Dale 1978; Kuss 1986; Wilson
and Seney 1994; Liddle 1997).
Off-Trail Impacts
Vegetation community composition
showed similar trends at both the species
and life form levels of analysis. This could
indicate that neither one, nor several, spe-
cies were driving community response.
Alternatively, a single/several species or
life form group could dominate across
large areas. In either case, generalities
can be made at the life form level. This
is advantageous since it means vegetation
community response can be considered at
a broader scale.
Life form community gradients did not
support hypotheses 2a and 2b. Forested
trails and dry trails showed sharper edges
(stronger species turnover gradients) than
open or wet trails. Dry forested trails
had the sharpest edge compared to the
other combinations of habitat/microhabitat.
Heath trails had the softest edge (lowest
amount of species turnover), whereas bog
trails had the sharpest edge (highest degree
of species turnover) compared to all other
habitat types.
Forest edges have been well documented
in the road ecology literature (Fraver 1994;
Murcia 1995; Ries et al. 2004; Marchand
and Houle 2006; Avon et al. 2010), although
road ecology studies have largely focused
on forests themselves (Fraver 1994; Murcia
1995; Ries et al. 2004; Marchand and Houle
2006; Avon et al. 2010), and studies that
did compare forests to open habitats looked
at relatively productive habitats such as
meadows or grasslands (Dale and Weaver
1974; Weaver and Dale 1978; Cole 1987).
Our study provides evidence of vegetation
community gradients extending beyond the
trail surface, particularly in low productiv-
ity open habitats. Trails in open habitats
show a softened edge compared to forest
trails, but have similar life form gradient
strengths as forest trails. This gradient is
most likely driven by bog sites rather than
heath sites since bog sites had the highest
community gradient and heath sites had
the weakest. The high community gradi-
ent on bog trails was expected since it
corresponds to past studies, and indicates
the high sensitivity of bog vegetation to
vehicular disturbance (Ross and Willison
1991; Charman and Pollard 1993).
Unexpectedly, the weakest community
gradient was found on heath trails. Else-
where, lichen dominated heath have been
found to be highly sensitive to on-trail
recreation impacts (i.e., direct trampling)
(Willard and Marr 1970; Greller et al. 1974;
Liddle 1997; Arnesen 1999). We predicted
dry heath would respond with a stronger
gradient, given that lichen abundance has
been documented to decrease near roads
(Glenn et al. 1993), likely due to their sen-
sitivity to air pollution (Ferry et al. 1973 as
cited in Angold 1997) and direct trampling
(Willard and Marr 1970; Bell and Bliss
1973). Perhaps stronger off-trail effects
were not found due to the low intensity
of traffic. Angold (1997) found that edge
effects in heath were strongly correlated to
the amount of traffic on the nearby road. In
our study, traffic levels were low, so pollu-
tion levels may be within tolerable limits
for lichen adjacent to the trail. This idea
of a “threshold” effect is also supported
by results found by Bayfield et al. (1981)
looking at the impact of walking paths on
lichen (Cladonia spp.) dominated heath in
Scotland. On lightly used paths, structural
damage to lichen beyond 1 m from the path
was low (Bayfield et al. 1981).
Graminoids showed strong edge associa-
tions in a number of habitat types, given
their high tolerance to trampling distur-
bance due to their morphological charac-
teristics of tough stems/tissues and basal
meristems (Cole 1995; Yorks et al. 1997),
which is consistent with findings of other
Habitat Type
Trees Shrubs Herbs Ferns Graminoid
Moss Lichen
Forest 0.330 -0.855 -0.455 -0.081 -0.486 -0.636 -0.661
Open 0.110 -0.382 -0.600 -0.231 -0.891 -0.477 0.382
Dry 0.127 -0.855 -0.455 -0.065 -0.771 -0.418 0.127
Wet -0.150 -0.818 -0.130 -0.085 -0.709 -0.745 -0.359
Forest/Dry 0.127 -0.917 -0.382 -0.070 -0.756 -0.673 -0.561
Forest/Wet -0.101 -0.964 -0.018 -0.185 -0.019 -0.208 0.370
Open/Dry 0.114 0.294 -0.164 -0.182 -0.278 0.330 0.382
Open/Wet 0.022 -0.527 -0.527 -0.217 -0.564 -0.881 -0.110
Kendall’s tau
Table 3. Correlation results of Polar Ordination life form analysis using Kendall’s tau to accommodate non-normality in the data. Negative correlation
indicates association with the edge. Correlations over 0.7 are bolded for emphasis. Data based on field work from June–September 2011 in and around
the Avalon Wilderness Reserve, island of Newfoundland, Canada.
276 Natural Areas Journal Volume 35 (2), 2015
researchers (Liddle and Greig-Smith 1975;
Hall and Kuss 1989). Mosses are another
group that showed strong edge associations
in a variety of habitat types. As a taxonomic
group, bryophytes are relatively tolerant
of trampling, with a notable exception
being members of the genus Sphagnum
(Studlar 1983; Cole 1995; Liddle 1997).
Morphological traits, such as small size
and compact growth, form convey tram-
pling resistance (Cole 1995; Yorks et al.
1997). Studlar (1983) noted that given suf-
ficient moisture, some species can exploit
disturbed ground. The finding that shrubs
had a strong association to the edge was
interesting since direct trampling studies
have indicated low resistance and resilience
to trampling in shrubs (Yorks et al. 1997);
in particular the chamaephytes (Cole 1995).
In this study, vegetation sampling began
directly beside the trail so that vegetation
was not directly trampled. Shrubs as a life
form group were able to exploit the nearby
disturbed, but untrampled, area. Lichens
were also weakly associated with the edge,
indicating the sensitivity of these two
groups to recreational disturbance. Lichen
sensitivity in alpine and heath environments
has been well recognized (Willard and Marr
1970; Bell and Bliss 1973; Bayfield et al.
1981); for example Bayfield et al. (1981)
found detectable small amounts of damage
to Cladonia sp. on heavily used walking
paths up to 50 m from the path (Bayfield
et al. 1981). Work by Cole (1995) suggests
that plant morphology is the most important
determinant of community resistance and
resilience for a single short-term trampling
event. However, as confirmed by our re-
sults, site (habitat) characteristics may be
important for areas that experience repeated
trampling events (Cole 1995).
On-Trail and Off-Trail Impacts:
Resistance and Resilience
All trails examined in this study received
some level of use. This study provides
valuable information about different habitat
responses under continuous recreational
use. This is important for management
decisions since recreational vehicle use is
unlikely to stop altogether within the wider
MBE. All habitat types were vulnerable to
either on-trail or off-trail impacts. There
was no one “super tolerant” community.
Broadly speaking, forested communities
were less resistant (strong edge), but more
resilient in drier stands (less erosion), com-
pared to heath and bog. Heath communities
were more resistant (softer edge), but less
resilient (more erosion). Bog communities
were neither resistant nor resilient.
A given habitat’s ability to sustain recre-
ational use depends upon its relative resis-
tance, resilience, or combination of those
two strategies—tolerance (Monz 2002).
Both plant morphological characteristics
and habitat characteristics (i.e., relative
productivity) play a role in determining
an ecosystem’s relative resistance, resil-
ience, and, ultimately, tolerance of human
recreational use. The boreal communi-
ties studied here may be relatively more
resistant to recreational vehicle erosion
compared to similar environments due
to the stoniness of the soils and lack of a
thick organic layer.
Previous ORV studies and vegetation
trampling studies in general have largely
focused on the trail surface or the area
immediately adjacent to it. This study
incorporates a gradient design to exam-
ine impacts into the interior of various
habitat types and within and outside a
protected area (intended as a surrogate for
traffic intensity). Based on our work, we
conclude that plant communities whose
tolerance to recreational vehicle impacts
is via high resistance and low resilience
could withstand periods of intense use.
However, once the impact threshold has
been reached (i.e., damage has occurred),
such communities would require periods
of recovery with no recreational use (Cole
1995; Gallet and Rozé 2001). Thus, man-
agers of such landscapes need to carefully
regulate the number of recreational users
to keep the level of use below the impact
threshold. Conversely, if managing for a
low resistance and high resilience sys-
tem, the impact threshold is likely to be
exceeded even at low levels of use. Here
regulating when recreationists use the
system is important (i.e., permitting lim-
ited seasonal use). While our work yields
management recommendations specific to
the Avalon Wilderness Reserve and wider
Maritime Barrens Ecoregion, we feel these
conclusions are applicable to similar low-
productivity, boreal environments.
This research was supported by a grant
from the Institute for Biodiversity, Eco-
system Science and Sustainability (IBES)
to YFW, and through in-kind field support
from the Newfoundland and Labrador De-
partment of Environment and Conservation
(Parks and Natural Areas Division and
Sustainable Development and Strategic
Science Division). We thank P. Howse for
assistance with field work, L. Morrissey
for help with plant identification, and L.
Siegwart Collier for her advice on multi-
variate statistical analysis. L. Hermanutz,
C. Purchase, and two anonymous reviewers
provided helpful comments on an earlier
version of the manuscript.
Nyssa Van Vierssen Trip was an MSc stu-
dent in Biology at Memorial University at
the time of this work. She has interests in
landscape and urban ecology. Since com-
pleting her MSc, she has relocated near
Ottawa, Canada, where she has been doing
consulting work for the Canadian Boreal
Forest Agreement and other clients. She is
continuing her studies as a PhD student
in the Faculty of Environmental Studies,
at York University, Toronto.
Yolanda Wiersma is a Landscape Ecologist
and Associate Professor in the Depart-
ment of Biology at Memorial University,
St. John’s, Newfoundland and Labrador,
Canada. She conducts interdisciplinary
research related to forestry, wildlife ecol-
ogy, and citizen science.
Allard, W. 2003. Detection of vegetation deg-
radation on Swedish mountainous heaths
at an early stage by image interpretation.
Ambio 32:510-519.
Angold, P.J. 1997. The impact of a road upon
adjacent heathland vegetation: effects on
plant species composition. Journal of Ap-
plied Ecology 34:409-417.
Arnesen, T. 1999. Vegetation dynamics follow-
Volume 35 (2), 2015 Natural Areas Journal 277
ing trampling in grassland and heathland in
Sølendet Nature Reserve, a boreal upland
area in central Norway. Nordic Journal of
Botany 19:47-69.
Avalon Wilderness Reserve Management Plan.
1986. Parks and Natural Areas Division,
Department of Environment and Conser-
vation Newfoundland and Labrador, St.
John’s, NL.
Avon, C., L. Bergès, Y. Dumas, and J.L. Dup-
ouey. 2010. Does the effect of forest roads
extend a few meters or more into the adjacent
forest? A study on understory plant diversity
in managed oak stands. Forest Ecology and
Management 259:1546-1555.
Bayfield, N.G. 1979. Recovery of four montane
health communities on Cairngorm, Scotland,
from disturbance by trampling. Biological
Conservation 15:165-179.
Bayfield, N.G., U.H. Urquhart, and S.M. Coo-
per. 1981. Susceptibility of four species of
Cladonia to disturbance by trampling in the
Cairngorm Mountains, Scotland. Journal of
Applied Ecology 18:303-310.
Beals, E.W. 1984. Bray-Curtis ordination: an
effective strategy for analysis of multivari-
ate ecological data. Advances in Ecological
Research 14:1-55.
Bell, K.L., and L.C. Bliss. 1973. Alpine distur-
bance studies: Olympic National Park, USA.
Biological Conservation 5:25-32.
Bellamy, D., J. Radwoth, and N.W. Radforth.
1971. Terrain, traffic and tundra. Nature
Bliss, L.C., and R.W. Wein. 1972. Plant com-
munity response to disturbances in the
western Canadian arctic. Canadian Journal
of Botany 50:1097-1109.
Brooks, M.L., and B. Lair. 2005. Ecological
effects of vehicular routes in a desert eco-
system. Technical Report, United States
Geological Survey, Western Ecological
Research Center, Las Vegas Field Station,
Henderson, NV.
Bryan, R.B. 1977. The influence of soil proper-
ties on degradation of mountain hiking trails
at Grövelsjön. Geografiska Annaler. Series
A. Physical Geography 59:49-65.
Buckley, R. 2004. Environmental impacts of
motorized off-highway vehicles. Pp. 83-99
in R. Buckley, ed., Environmental Impacts
of Ecotourism. CABI Publishing, Welling-
ford, UK.
Burnham, K.P., and D.R. Anderson. 2002. Mod-
el Selection and Multimodel Inference: A
Practical Information-Theoretic Approach,
2nd ed. Springer, New York.
Burton, R.C.J. 1974. The recreational carrying
capacity of the countryside. Occasional Pub-
lication No. 11, Keele University Library,
Keele, UK.
Charman, D.J., and A.J. Pollard. 1993. Long
term vegetation recovery after vehicle
track abandonment on Dartmoor, south-
west England, UK. The British Ecological
Society 25:22-28.
Clarke, K.R. 1993. Non-parametric multi-
variate analyses of changes in community
structure. Australian Journal of Ecology
Cliff, A.D., and J.K. Ord. 1981. Spatial
processes: models and applications. Pion
Limited, London, UK.
Cole, D.N. 1978. Estimating the susceptibil-
ity of wildland vegetation to trailside
alteration. Journal of Applied of Ecology
Cole, D.N. 1987. Effects of three seasons of
experimental trampling on five montane for-
est communities and a grassland in western
Montana, USA. Biological Conversation
Cole, D.N. 1992. Modeling wilderness
campsites: factors that influence amount
of impact. Environmental Management
Cole, D.N. 1995. Experimental trampling of
vegetation, II. Predictors of resistance and
resilience. Journal of Applied Ecology
Cole D.N. 2004. Impacts of hiking and camp-
ing on soils and vegetation: a review. Pp.
41-61 in R. Buckley, ed., Environmental
Impacts of Ecotourism. CABI Publishing,
Wellingford, UK.
Dale, D., and T. Weaver. 1974. Trampling effects
on vegetation of the trail corridors of north
rocky mountain forests. Journal of Applied
Ecology 11:767-772.
Damman, A.W.H. 1983. An ecological subdi-
vision of the island of Newfoundland. Pp.
163-206 in G.R. South, ed., Biogeography
and Ecology of the Island of Newfoundland.
Dr. W. Junk Publishers, The Hague.
Ferry, B.W., H.S. Baddeley, and D.L. Hawkes-
worth. 1973. Air Pollution and Lichens.
Athlone Press, London.
Fortin, M.J., and M. Dale. 2005. Spatial
analysis: a guide for ecologists. Cambridge
University Press, UK.
Fraver, S. 1994. Vegetation responses along
edge-to-interior gradients in mixed hard-
wood forests of the Roanoke River basin,
North Carolina. Conservation Biology
Gallet, S., and F. Rozé. 2001. Resistance of
Atlantic healthlands to trampling in Brittney
(France): influence of vegetation type,
season and weather conditions. Biological
Conservation 97:189-198.
Glenn, M., A. Gomez-Bolea, C. Ropux, and
R. Labella. 1993. Lichen heath in relation
to vehicular traffic in Mediterranean forest
reserves (abstract only). American Journal
of Botany 80 (supplemental) 6:8.
Greller, A.W., M. Goldstein, and L. Marcus.
1974. Snowmobile impact on three alpine
tundra plant communities. Environmental
Conservation 1:101-110.
Hall, C.H., and F.R. Kuss. 1989. Vegetation al-
teration along trails in Shenandoah National
Park, Virginia. Biological Conservation 48:
Iverson, R.M., B.S. Hinckley, R.M. Webb, and
B. Hallet. 1981. Physical effects of vehicular
disturbance on arid landscapes. Science
212: 915-917.
Jones, D.H. 1978. The effect of pedestrian im-
pact on selected soils. MS thesis. University
of Glasgow, Scotland.
Jorgenson, J.C., J.M. Ver Hoef, and M.T. Jor-
genson. 2010. Long-term recovery patterns
of arctic tundra after winter seismic explora-
tion. Ecological Applications 20:205-221.
Kelleway, J. 2005. Ecological impacts of rec-
reational vehicle use on saltmarshes of the
Georges River, Sydney [Australia]. Wetlands
Kuss, F.R. 1986. A review of major factors
influencing plant responses to recreation
impacts. Environmental Management
Kuss, F.R., and C.N. Hall. 1991. Ground flora
trampling studies: five years after closure.
Environmental Management 15:715-727.
Lagocki, H.F.R. 1978. Surface soil stability as
defined and controlled by a drainage crite-
rion. Pp. 233-237 in W.W. Emerson, R.D.
Bond, and A.R. Dexter, eds., Modification of
Soil Structure. John Wiley, Hoboken, NJ.
Legg, M. 1973. Site factors useful in predicting
deterioration of forest campsites in northern
Michigan. PhD diss., Michigan State Uni-
versity, East Lansing.
Leung, Y.F., and J.L. Marion. 2000. Recreation
impacts and management in wilderness: A
state-of-knowledge Review. USDA Forest
Service Proceedings 5:23-48.
Liddle, M.J. 1975. A theoretical relationship
between the primary productivity of veg-
etation and its ability to tolerate trampling.
Biological Conservation 8:251-256.
Liddle, M.J. 1997. Recreation Ecology: The
Ecological Impact of Outdoor Recreation
and Ecotourism. Chapman & Hall, New
Liddle, M.J., and P. Grieg-Smith. 1975. A
survey of tracks and paths in a sand dune
ecosystem, II. Vegetation. Journal of Applied
Ecology 12:909-930.
Marchand, P., and G. Houle. 2006. Spatial
278 Natural Areas Journal Volume 35 (2), 2015
patterns of plant species richness along a
forest edge: What are their determinants?
Forest and Ecology and Management
Marshall, T.J., and J.W. Holmes. 1979. Soil
Physics. Cambridge University Press, UK.
McCune, B., and J.B. Grace. 2002. Analysis of
ecological communities with a contribution
from D.L. Urban. MjM Software Design,
Gleneden Beach, OR.
McCune, B., and M.J. Mefford. 2011. PC-ORD.
Multivariate analysis of ecological data. Ver-
sion 6.0. MjM, Gleneden Beach, OR.
Meadows, D., R. Foltz, and N. Geehan. 2008.
Effects of all-terrain vehicles on forested
lands and grasslands. 0823 1811-SDTDC,
US Department of Agriculture, Forest Ser-
vice, San Dimas Technology and Develop-
ment Center, San Dimas, CA.
Mielke, P.W., Jr. 1984. Meteorological appli-
cations of permutations techniques based
on distance functions. Pp. 813-830 in P.R.
Krishnaiah, and P.K. Sen, eds., Handbook of
Statistics, Vol. 4. Elsevier Science Publish-
ers, Amsterdam.
Mielke, P.W., Jr., and K.J. Berry. 2007. Per-
mutation Methods: A Distance Function
Approach, 2nd ed. Springer, New York.
Minchin, P.R. 1987. An evaluation of the rela-
tive robustness of techniques for ecological
ordination. Vegetatio 69:89-107.
Monz, C.A. 2002. The response of two arctic
tundra plant communities to human tram-
pling disturbance. Journal of Environmental
Management 64:207-217.
Monz, C.A., C.M. Pickering, and W.L. Had-
wen. 2013. Recent advances in recreation
ecology and the implications of different
relationships between recreation use and
ecological impacts. Frontiers in Ecology
and the Environment 11:441-446.
Murcia, C. 1995. Edge effects in fragmented for-
ests: implications for conservation. Trends
in Ecology and Evolution 10:58-62.
Olive, N.D., and J.L. Marion. 2009. The
influence of use-related, environmental,
and managerial factors on soil loss from
recreational trails. Journal of Environmental
Management 90:1483-1493.
Ouren, D.S., C. Haas, C.P. Melcher, S.C.
Stewart, P.D. Ponds, N.R. Sexton, L. Burris,
T. Fencher, and Z.H. Bowen. 2007. Envi-
ronmental effects of off-highway vehicles
on Bureau of Land Management lands: a
literature synthesis, annotated bibliogra-
phies, extensive bibliographies, and internet
resources: Open-File Report 2007-1353,
United States Geological Survey, Fort Col-
lins Science Center, CO.
Peck, J.E. 2010. Multivariate analysis for
community ecologists: step-by-step using
PC-ORD. MjM Software Design, Gleneden
Beach, OR.
Pickering, C.M., and W. Hill. 2007. Impacts of
recreation and tourism on plant biodiversity
and vegetation in protected areas in Austra-
lia. Journal of Environmental Management
Quinn, M., and G. Chernoff. 2010. Mountain
biking: a review of the ecological effects. A
literature review for Parks Canada- National
Office (Visitor Experience Branch). Final
Report, February 2010, Ottawa, ON.
R Core Team. 2011. R: A language and environ-
ment for statistical computing. Version 14.1.
The R Project for Statistical Computing.
Radforth, J.R. 1972. Analysis of disturbance
effects of operations of off-road vehicles on
tundra. ALUR contract report, Department
of Indian Affairs and Northern Development,
Ottawa, ON.
Ries, L., R.J. Fletcher Jr., J. Battin, and T.D.
Sisk. 2004. Ecological responses to habitat
models: mechanisms, models, and variability
explained. Annual Review of Ecology, Evo-
lution and Systematics 35:491-522.
Roberts, B.A. 1983. Soils. Pp. 107-162 in G.R.
South, ed., Biogeography and Ecology of
the Island of Newfoundland. Dr. W. Junk
Publishers, The Hague.
Rooney, T.P. 2005. Distribution of ecologi-
cally-invasive plants along off-road vehicle
trails in the Chequamegon National Forest,
Wisconsin. The Michigan Botanist 44:178-
Ross, J.B. , and J.H.M. Willison. 1991. Impacts
of all-terrain vehicles on bogs and barrens of
the Cape Breton Highlands. Pp. 533-534 in
Proceedings of the International Conference
on Science and Management of Protected
Areas, 14-19 May 1991. Halifax, Nova
Scotia, Canada.
Sokal, R.R., and N.L. Oden. 1978. Spatial
autocorrelation in biology 1. Methodology.
Biological Journal of the Linnean Society
Studlar, S.M. 1983. Recovery of trampling
bryophyte communities near Mountain Lake,
Virginia. Bulletin of the Torrey Botanical
Club 110:1-11.
Törn, A., A. Tolvanen, Y. Norokorpi, R. Tervo,
and P. Siikamäki. 2009. Comparing the
impacts of hiking, skiing, and horse riding
on trail and vegetation in different types of
forest. Journal of Environmental Manage-
ment 90:1427-1434.
van Vierssen Trip, N. 2014. A Comparison of
All-Terrain Vehicle (ATV) Trail Impacts
in Boreal Forest, Heath and Bog Habitats
within the Avalon Wilderness Reserve and
Surrounding Area. MSc thesis, Department
of Biology, Memorial University, St. John’s,
NL, Canada.
Weaver, T., and D. Dale. 1978. Trampling ef-
fects of hiker, motorcycles and horses in
meadows and forests. Journal of Applied
Ecology 15:451-457.
Willard, B.E., and J.W. Marr. 1970. Effects of
human activities on alpine tundra ecosystems
in Rocky Mountain National Park. Biologi-
cal Conservation 2:257-265.
Wilson, J.P., and J.P. Seney. 1994. Erosional
impact of hikers, horses, motorcycles, and
off-road bicycles on mountain trails in
Montana. Mountain Research and Develop-
ment 14:77-88.
Yorks, T.P., N.E. West, R.J. Mueller, and S.D.
Warren. 1997. Toleration of traffic by veg-
etation: life form conclusions and summary
extracts from a comprehensive data base.
Environmental Management 21:121-131.
... In the US, where most of the research was conducted, ORVs are state regulated, making it difficult to determine the exact extent of ORV activities at a national scale. While the US is thought to have one of the highest rates of ORV activity, off-road driving is altering landscapes across Europe (Enríquez-de-Salamanca, 2021), Australia (Davies et al., 2016), Canada (Trip and Wiersma, 2015), as well as less-developed nations like Mongolia (Li et al., 2006), Argentina (Navas Romero et al., 2019), and Saudi Arabia (Assaeed et al., 2019). ...
... Interestingly, in both a desert and grassland habitat, ORV activity altered the particle size of the soil, increasing the percentage of coarse particles compared to finer particles (Pérez, 1991;Brown and Schoknecht, 2001;Al-Awadhi, 2013). Several studies reported significant ruts being created in designated ORV areas in desert (Webb et al., 2013), coastal (Davies et al., 2016) and grassland habitats (Trip and Wiersma, 2015). The depth of rutting may be attributed to increasing vehicle passes (Webb et al., 2013) or habitat type (Trip and Wiersma, 2015), but factors such as seasonal changes in soil moisture have been largely understudied. ...
... Several studies reported significant ruts being created in designated ORV areas in desert (Webb et al., 2013), coastal (Davies et al., 2016) and grassland habitats (Trip and Wiersma, 2015). The depth of rutting may be attributed to increasing vehicle passes (Webb et al., 2013) or habitat type (Trip and Wiersma, 2015), but factors such as seasonal changes in soil moisture have been largely understudied. For example, in a study comparing the impacts of ORV use on grassland, forest, and wetland habitats, rut depth was deeper on heath and bog trails compared to forested trails and on wet soils compared to dry (Trip and Wiersma, 2015). ...
Full-text available
The global use of off-road vehicles (ORVs) in natural environments has accelerated rapidly over the last few decades, resulting in significant social and environmental consequences. As the demand, use, and promotion of light-duty ORVs like all-terrain vehicles (ATVs), motorcycles, four-wheel drive trucks and sport utility vehicles (SUVs) increases in remote wilderness, the landscape is becoming fragmented into disorganized and destructive networks of trails and roads. Substantial ecological impacts to a wide range of ecosystem structures and functions will likely result from ORV activity. Applying a global systematic review, we examine 105 publications about plant, soil, and wildlife responses to ORV traffic in different habitats to help guide the direction of future research, monitoring programs, and mitigation efforts. Most studies investigated impacts to animals, followed by soils, then vegetative responses. Soil studies primarily focused on physical impacts to the soil (i.e., compaction, erosion, rut depth), but some studies suggest that soil chemical and biological properties may also be impacted by ORV traffic. The literature on plant responses to ORV activities primarily explored vegetation loss, although impacts on the plant community were also investigated. Animal studies investigated impacts of ORV use on invertebrates, mammals, birds, and to a lesser extent reptiles/amphibians, including population-level, community-level, and behavioral responses. Overall, research on environmental impacts of ORV traffic is biased to coastal and desert ecosystems in the northern hemisphere (primarily in the US), often does not address mechanisms that may produce ecological impacts (e.g., intensity of vehicular disturbance and ecosystem- or species-specific sensitivity to ORV activities), and frequently focused on short-term responses. More research is needed to understand the mechanisms that cause the different responses of soil, plant, and animals to ORVs over the long-term in a broad range of ecosystems to support real-time management and conservation efforts.
... Thus, large quartzite rocks were used to fulfil the huge gully erosion and to reconstruct and pave the surface (Fig. 2 a-III, b-III) to re-direct the water flow to its original paths (Akbarimehr and Naghdi, 2012;Yitbarek et al., 2012). Besides rehabilitating the water flow path in ES-Erosion C, a suspended bridge was constructed to protect the water flow system and allow access to trail 2, while trail 1 was inactivated and closed ( Fig. 2 c-IV), avoiding disturbances from trail users (Trip and Wiersma, 2015). Indeed, trail closure is an effective strategy to reduce erosion and, the construction of a suspended bridge may be a useful solution to protect these trails (Barros et al., 2013;Tomczyk et al., 2016). ...
The quartzite rock outcrops and the native vegetation of grasslands located at the Serra da Calçada Mountain in Minas Gerais State (Brazil) have been severely degraded by extreme sports activities such as motocross and off-road vehicles, greatly damaging the abundant headwaters. The main consequences thereof were hilly and gully erosion processes with soil loss and the deviation of the water from its original paths. However, currently, there is no report of successful restoration efforts in severely eroded outcrops in Brazilian high-altitude grasslands (campo rupestre). Through the Universal Soil Loss Equation (USLE), we found a high general erosion rate in the study site (669.91 t·ha⁻¹·year⁻¹), and the specific soil loss provoked by off-road vehicles on trails was significantly greater (49 m³ per 100 m²) than that caused by mountain bikes and trekking (5.8 m³ per 100 m²). We performed the physical reconstruction of eroded outcrops and surface water flow paths by allocating locally available quartzite rocks. These rocks were inoculated with different species of bryophytes and planted with native species under two treatments: un-inoculated and inoculated with arbuscular mycorrhizal fungi (AMF) spores of the Rhizophagus irregularis species. After 2 years, the bryophyte communities showed a similar pattern to the preserved site, and the AMF inoculation favoured plant establishment of most species, especially of the Asteraceae, Cyperaceae, Fabaceae, Malpighiaceae, Orchidaceae and Poaceae families. The AMF also improved the soil fertility, highlighting soil P, SOM, CEC, NH4⁺-N as well as soil water content and water retention capacity. Poaceae family species showed an outstanding occupation, which was considered a functional indicator of rehabilitation success, functioning as a “hydraulic carpet” for water exportation, conduction and drainage across the outcrops. This study provides an eco-technology to restore severely eroded outcrops over headwaters using native species in the Brazilian high-altitude grasslands.
... Point counts indicated that Canada Warblers avoided newly created seismic lines or seismic lines that were regularly used by recreational or industrial off-highway vehicles (OHVs). Regular use of OHVs in the boreal forest can impede regeneration on the disturbance and often leads to seismic lines becoming wider over time as operators avoid wet areas, causing damage to the adjacent forest vegetation (Thurston andReader 2001, van Vierssen Trip andWiersma 2015). This also means that Canada Warbler avoidance of mostly disturbed lines may not be driven solely by the lack of vegetation on the line. ...
... Findings also show an interesting behavior of the vegetation cover on test field 2 where, after at least 2 years without or with minor disturbance, plants started taking roots in the central area of the field marking the final part of the recovery phase. The topic of plant's resilience to trampling and vehicle movement was a topic of other studies [37], [38], [39], where it was confirmed their importance in the ability of the natural route restoration. As it currently stands, results of both test fields show that the route natural restoration ability is not enough as the devastation cycle is too often repeated on the forested roads in Kielce. ...
Conference Paper
Full-text available
Usage of the tourist routes by the tourists causes a number of negative consequences which ranges from soil loss by rill and sheet erosion, trampling, increased compaction, and vegetation destruction to a disturbance in water retention and infiltration. In the close proximity of the cities, economic and recreational interests meet in the form of tourist routes that usually are designated on already existing roads, throughout the forests used for heavy vehicles movement that are not under active environmental protection. The main goal of this paper is to determine micro-topographical consequences of heavy vehicle passages and a tourist routes surfaces changes with in 10 year period in such area. Fruition of this goal requires monitoring procedures that allows for capturing changes over time. For that purpose, LiDAR data and the Structure-from-Motion photogrammetry technique has been applied in two separated plots on tourist routes surfaces affected by vehicles transitions in order to obtain detailed digital elevation models. Those models, in turn, allow for calculation and analysis of the changes over time with the use of digital elevation model of difference method. After two years of monitoring which includes total of twelve scans for both test fields, models were then analyzed to draw conclusions. Results of observations and scanning show destructive effects caused by the vehicles passages as well as further stages of the surface evolution including initial recovery stage and main transformation stage.
... Local community-established protected areas have the potential to advance large-scale conservation goals, such as goals tied to biodiversity protection, even if these goals may not have motivated local protection (Kroetz et al. 2014;Crain et al. 2020;Stachowiak et al. 2021). However, the extent to which this potential can be achieved will depend on the ecological condition of these local community-established sites, and ecological conditions on protected areas can vary greatly depending on how sites are used (Reed and Merenlender 2008;Hull et al. 2014;Trip and Wiersma, 2015) and managed (Leverington et al. 2010;Kearney et al. 2020). ...
Full-text available
Local communities often conserve nearby natural areas to support recreational activities and other benefits these areas provide. Areas protected by local communities could contribute to wider efforts to achieve large-scale conservation goals, such as biodiversity protection, provided the ecological conditions on-site are compatible with achieving these goals. To explore the potential contribution of locally established protected areas, we focus on areas protected by local communities in California, USA, using ballot initiatives, a form of direct democracy. We compare the ecological condition of wooded habitat on protected areas funded by local communities through the ballot box to that of similar habitats on protected areas funded by a state conservation agency. As an indicator of ecological condition, we focus on coverage by exotic plant species. We examine whether protected area characteristics or aspects of human-mediated onsite disturbance related to recreational use explain exotic plant cover found on each type of protected area. Exotic plant cover did not differ between areas protected by local communities and those protected by our larger scale conservation actor. Instead, elevation was the best predictor of exotic plant cover. Our results suggest protected areas established by local communities may be in no worse a condition than those established by a state public agency and warrant inclusion when tracking progress towards large-scale conservation goals for protected areas.
... These impacts can diminish the attractiveness, desirability, and functionality of recreational areas and facilities (Cole 1986, Cole 1996, Monz et al. 2010). Susceptibility to direct and indirect disturbances varies across different ecosystems and across types of use (Trip & Wiersma 2015, Havlick et al. 2016. ...
Full-text available
Human recreational and agricultural activities within public forests generate unintended perturbations for wildlife occupying the forests. Many of these effects likely will persist in perpetuity and need study. I inventoried the wild birds and mammals occupying a multiuse National Forest in southeastern Idaho, USA by measuring their presence through the use of remote sensing camera traps, avian abundance counts, and small mammal live-trapping. Simultaneously, I measured human use of trails and occurrence of livestock on trails within the forest through camera traps placed throughout the public trail network in the forest. I statistically modeled relationships between types of human recreation or presence of livestock and the abundance of wildlife from measured taxa, i.e. medium- and large- bodied mammals, small mammals, and birds. Wildlife did not exhibit universal or ubiquitous responses at the community level to human or livestock use. Rather, I observed significant taxon-specific responses. For example, mule deer (Odocoileus hemionus) were highly responsive to motorized use of public trails, exhibiting a pronounced temporal avoidance, whereas coyotes (Canis latrans) predictably co-occurred with presence of domestic cattle (Bos taurus). As humans increase their use of public forests for recreation and continue agricultural use, it is imperative for land and wildlife managers to understand how human activities within public forests influence the wildlife species occupying these forests in order to conserve forest wildlife communities effectively.
Recreational activities are among the most common threats to species at risk. The use of standardized threat assessment tables in assessment reports for 300 Canadian species at risk since 2012 enabled a systematic comparison of threat categories considering both frequency and intensity of threats. Our analysis of these reports reveals that recreational activities are the most common threat to species at risk in Canada, affecting more species than any other threat category. However, the intensity of a threat should be considered. When accounting for the intensity of threats, recreational activities are the third-greatest threat to species at risk in Canada following “Invasive Non-native/Alien Species” and “Roads and Railroads”. Recreational activities were among the top five threats to molluscs, vascular plants, mosses, arthropods, marine and terrestrial mammals, and reptiles considering both frequency and intensity of threat to the assessed species in each taxonomic group. Among species at risk for which recreational activities posed at least a low-level threat, off-road vehicle use was the most commonly mentioned recreational threat. The second-most common recreational threat was hiking. Boating, mountain biking, camping, beach use, horseback riding, off-leash pets, and rock climbing were considered at least a low-level threat to five or more species. Common mechanisms of threat arising from recreational activities include direct mortality and damage from trampling species and/or their habitat, siltation and changes to water chemistry, disruption due to human presence, and noise disturbance. Management implications Recreation activities affected more species at risk than any other category of threat. The intensity of threat from recreational activities was negligible or low for most at-risk species affected. However, increases in recreational use and cumulative effects could result in more severe threats in the future. Both mechanized and non-mechanized recreational activities cause direct mortality, disturbance, and/or degradation of habitat for species at risk in Canada including many under-studied taxa. Thus, recreational activities that are typically unauthorised and those that are commonly permitted in protected areas warrant management attention to avoid impacts on species at risk, especially less conspicuous taxa.
Full-text available
Enhancing the resilience of the forests has become increasingly more important as climate changes and disturbances become more severe and frequent. Disturbances such as drought, bark beetle and fires are causing excessive tree mortality across Europe. Resilience is seen as an answer to improve the capacity of forests to persist despite disturbances and adapt to climate change. New policies demand measures to increase forest resilience to ensure the provision of crucial ecosystem services to facilitate the transition from fossil-based to bio-based economy. However, resilience is a debated concept with multiple definitions in science. The variety of definitions has led to a lack of common metrics for measurement. The unclarity in how to define and measure resilience makes it difficult for forest owners and managers to implement the concept in their forest management. There is a dire need to make resilience more operationalised to help the forests to better cope with climate change and the increased disturbances. In this research, we addressed the knowledge gap in how to operationalise resilience in forest management by providing for a frame for navigating the different definitions and giving examples on how they can be measured. To do this, we first reviewed how resilience is used in forest sciences in terms of definition and measurements. In the literature, three main resilience concepts dominate: engineering resilience (“recovery of a previous state”), ecological resilience (“remaining within the prevailing system domain through maintaining important ecosystem processes and functions”) and social-ecological resilience (“the capacity to reorganize and adapt through multi-scale interactions between social and ecological components of the system). We examined how similar the three concepts were by analysing the types of research settings these three definitions were used, how they were assessed, and what indicators were used. Next, we developed a Principle, Criteria, and Indicator -framework to help forest managers to identify how forest management objectives and trade-offs influence resilience and how the trade-offs could be balanced to achieve more resilient forest. In addition, we analysed the use of high-resolution dendrometers as a tool for monitoring tree stress and resilience to drought. Finally, we explored how the science-practice interphase in forest management could be improved to facilitate the transferring of the scientific knowledge on how to improve resilience to disturbances into practical forest management. We found that the three common concepts of resilience are not contrasting but instead they are complimentary to one another and form a nested hierarchy where engineering resilience is nested inside ecological resilience, which in turn is nested inside the social-ecological resilience. Their use depends on the complexity of the researched system with engineering resilience used for simple systems and ecological and social-ecological resilience for more complex ones. Therefore, instead of debating on the correct definition to use, forest managers should carefully determine of which part of the forest or forest value chain they are managing, to what they need to increase the resilience to, and who are likely to be influenced by their decisions. We were furthermore able to show with the developed framework that the forest management goals influence the trade-offs in forest management and the level of resilience of forest and the surrounding society, indicating that the steps to achieve resilient forest differ depending on the management goal. In addition, the results showed that high-resolution dendrometers have the potential to inform forest managers on the stress and resilience of trees to dry conditions, however more research is still needed before the tool can be used in the practical management level. Lastly, we found that while the science-practice interphase is valued by the forest professionals, there is in some cases weak evidence for the effectiveness of the forest management measures proposed by forest professionals. Moreover, many forest professionals face considerable barriers in implementing resilience into forest management. To conclude, the research we conducted provided remarkable advances on operationalising resilience into forest management. Our results showed that resilience can be implemented into forest management with a variety of forest management goals and strategies. The future research should focus on developing, together with practitioners, resilience indicators for forests under different management regimes across Europe. Moreover, efforts to study the impacts of different forest management measures on resilience to disturbances should be increased. However, forest-related policies and management practices should already proceed to incorporate measures to enhance resilience of forests to ensure the provisioning of ecosystem services.
Saltmarshes provide multiple ecosystem services, and some have been preserved as conservation areas. Studies indicate recreation-related vehicle use may be significantly degrading them. Saltmarshes have a low resistance and resilience to recreation-related trampling impacts; however, little is known about those associated with vehicle use under a specific management strategy. Drone imagery and GIS spatial analysis were used to determine the area and intensity of direct vehicle impacts within a New Zealand saltmarsh. The management plan allows vehicle entry as long as it does not substantially impact cultural, ecological, or mahinga kai (or food cultivation) values. It limits impacts of vehicles by meeting two goals: limiting the area of vehicle use to formal road corridors, and limiting particularly damaging behaviors of use, including entering the saltmarsh when conditions are wet, traveling more than 10 km per hour, and using traction equipment. Results demonstrated substantial impacts. Tire tracks were present in 66% of quadrats sampled, and were distributed across the length and breadth of the saltmarsh, covering 17% (approximately 207 ha) of the 1225 ha saltmarsh. About a third of these quadrats had track covers of 1150%. Furthermore, particularly damaging vehicle use behaviors were widely evidenced, including deeply rutted mud, water channel initiation, and substantial loss and fragmentation of vegetation communities. Vehicle use is clearly eroding at least some of the cultural, ecological, and/or mahinga kai values for which the saltmarsh was conserved. While there are many indirect and direct measures for improving the current management strategy, none is likely to result in substantial reductions in vehicle impacts given the low resistance and resilience of saltmarshes to trampling. An alternative strategy that would lead to substantial reductions, and eventual recovery of the saltmarsh, would be to allow existing recreational activities, but deny vehicle entry.
Questions Increases in all‐terrain vehicle (ATV) use on dunes raises concerns about an ecosystem vital for coastal protection. We asked: with distance from trails, what are the effects of ATV use on (a) total, native, and non‐native plant species richness and (b) presence and cover of the dune‐stabilising plant Ammophila breviligulata? Specifically, how do (a) and (b) differ (1) between regions with and without ATV use; (2) with deeper ruts and increased distance from the ATV trail; and (3) between pioneer and shrub zones of dunes in each region? Location Miscou Island and Kouchibouguac National Park, New Brunswick, Canada. Methods We assessed ATV effects by conducting field vegetation surveys in a region with (Miscou Island) and without (Kouchibouguac National Park) ATV use. Line transects were used to capture gradients of effects across the dune community via plots evenly placed to measure trail effects (on the trail), close‐edge effects (edge of the trail), and distant‐edge effects (every 5 m up to 25 m away from trail), in pioneer and shrub zones of dunes. Results All‐terrain vehicle rut depth was associated with a decrease in total and native species on the trails and on the edge of trails, and with a slight increase in non‐native species beyond the trail edge. We also found a rut depth threshold of approximately 50 cm, beyond which was an abrupt decline across all species. Where ATV activity occurred, there was also a decrease of A. breviligulata in presence and cover, non‐native species increased in the pioneer zone, and the shrub zone had fewer native species. Conclusions All‐terrain vehicle use plays a major role in the vegetation changes observed on coastal dunes. A management plan that recognises the specific effects caused by ATV use on dune vegetation will help preserve dunes, enabling more cost‐effective coastal protection than engineered interventions.
The ecoregion is homologous with the major climatic subregions of Newfoundland. Bearing in mind site differences related to such factors as topography and history of disturbance, each ecoregion has a distinctive, recurring pattern of vegetation and soil development, controlled by regional climate. The general characteristics, vegetation, climate and (where appropriate) major geographical variation, effects of altitude, changes in lithology, climatic gradients and intra-ecoregional subdivisions are described for the following ecoregions: Central Newfoundland, Northern Peninsula, Avalon Forest, Maritime Barrens Eastern Hyper-oceanic Barrens, Long Range Barrens (highlands from the SW coast to the northern part of the Northern Peninsula), and Strait of Belle Isle. The phytogeographical position of these ecoregions is discussed.-P.J.Jarvis
The number and variety of statistical techniques for spatial analysis of ecological data are burgeoning and many ecologists are unfamiliar with what is available and how the techniques should be used. This book provides an overview of the wide range of spatial statistics available to analyze ecological data, and provides advice and guidance for graduate students and practicing researchers who are either about to embark on spatial analysis in ecological studies or who have started but need guidance to proceed. © M.-J. Fortin and M.R.T. Dale 2005 and Cambridge University Press 2009.
The heavy vehicles of the oilmen roam widely over Northern Canada, leaving trails gouged out of the virgin tundra. The damage they inflict on the vegetation may mark them as a basic problem in the conservation of the Arctic.
Publisher Summary Ordination implies an abstract space in which the entities form a constellation. In the Bray–Curtis ordination, the entities are samples and the attributes are species values in those samples. The aim of this method is to (1) calculate a distance matrix, (2) select two reference points (either real or synthetic samples) for determining direction of each axis, and (3) project all samples onto each such axis by their relationship to the two reference points. There are two major problems common to all ordination techniques, which include a function of the β-diversity or heterogeneity of the data set—that is, how different the samples are from one another. All ordinations distort the original multivariate data set and information is inevitably lost. Distortion in ordination has two kinds of consequences. The first is compressing and stretching distances in the ordination, compared with the original distance measures and relative to one another. The second consequence is the curvature of environmental axes, and this relates to Orloci's types A and C. Some of the alternatives to Bray–Curtis ordination are principal component analysis, reciprocal averaging, and iterative-stress minimization techniques.
(1) Damage to Cladonia uncialis, C. arbuscula, C. rangiferina and C. impexa resulting from human trampling was recorded in kilometre squares with levels of visitor use from 0 to > 18.8 people day$^{-1}$ (2) Damage declined with distance from paths, and increased with level of use. There was generally little damage 50 m from paths, except on the most heavily used sites. (3) Damage was more extensive at sites which were `open' in character than at those where use was `confined' (4) All four species changed from a pliable to a brittle state at water contents below c. 25%. (5) There is probably widespread slight damage on the `open' plateau areas of the Cairngorms receiving moderate to heavy visitor use, as well as more serious disturbance of lichens adjacent to paths.
Controlled levels of human trampling fragmented and, after one year, rejuvenated six species of mosses. Polytrichum commune, Ditrichum pallidum, Thuidium delicatulum, Hypnum imponens and Sphagnum palustre/henryense showed c. 80% or more recovery from 4200 walks. Sphagnum recurvum showed 98 and 25% recovery from 130 and 1600 walks, respectively. After two years, Polytrichum and S. recurvum showed further recovery while Ditrichum abruptly declined. Recovery rates varied widely within species and were affected by initial colony size, microtopography, litter levels, and competition from other bryophyte species and vascular plants.