Outdoor Recreation Activities as Vectors
for Invasive Species in the Forest
David Lisle

Abstract
Invasive species are generally regarded as a serious threat to forest ecosystems, however, the role
that outdoor recreational activity plays in their spread remains understudied. This study
examined whether the density of invasive plant species varied by distance from recreation trails
categorized by high and low usage intensity. Strava global heat maps were used to determine
high and low usage trails, and plotted perspective points on the Gaia GPS app. At each sampling
location, I counted the number of invasive plants in two quadrats, one close and one far from the
trail. The results showed no significant difference in invasive species density between high-use
and low-use trails. Further, there was no significant difference in the density of invasive species
close to and far from both types of trails. These findings suggest that humans may not have a
significant role in the spread of invasive species along trails, and there may be other factors that
affect the distribution in densities of invasive plants. Despite this, the findings support the need
for more public awareness, and targeted management strategies in areas that are the most
affected.
Key words: Invasive species, Outdoor Recreation, Distance from Trails, Usage Intensity,
Density.

Introduction
Invasive plants outcompeting native flora is a well-studied phenomenon. In the past thirty years,
research on the threats of invasive species to ecosystems, economies, and human health has been
increasing (Zimmerman et al., 2014). Invasive plants have negative consequences for ecosystems
because they decrease biodiversity, reduce trophic productivity, outcompete native vegetation,
and alter ecosystem structure (Davies, 2011). Davies et al., (2011) found that native vegetation
cover decreased with increasing density of an invasive grass, leading to decreased species
richness, and diversity. In addition, invasive species may change nutrient pools, and alter natural
fire regimes (Barney et al., 2013). Schmitz and Simberloff (1997) found that exotic (invasive)
species have played a role in 42% of the declines of endangered and threatened species in the
USA. Despite these well-studied impacts, non-native plants and seeds continue to be sold by
garden centers or plant nurseries, therefore adding to a problem experienced around the world
(Darcy & Burkart, 2002).

1

Recent studies have shown that the spread of invasive species is primarily caused by humans,
through a variety of methods such as roads, and recreation trails (Christen & Matlack, 2009;
Liedtke et al., 2020; Aziz et al., 2023). Once established, invasive species spread easily due to
their dispersal ability (Coutts et al., 2010) Invasive plants often have higher population growth
rates and greater reproductive capacities than native plants (Coutts et al., 2010). To accomplish
rapid spread, invasive plants often have life cycle strategies that focus on rapid reproduction
(Sakai et al., 2001). For example, some invasive plants can reproduce both asexually (cloning),
and sexually (seeds, spores) to accomplish rapid spread (Sakai et al., 2001). Furthermore, the
biggest environmental dispersers of invasive seeds are wind, water, and animals (mainly birds)
(Sakai et al., 2001). Humans, however, have recently surpassed environmental dispersal,
becoming the biggest spreader of invasive plants globally (Mack and Lonsdale, 2001). Humans
disperse invasive species via roads, outdoor recreation, and their pets (Mack and Lonsdale,
2001). Roads act as both habitat, and conduits for the spread of invasive plants (Christen &
Matlack, 2009). Human disruption of natural areas can cause negative impacts for ecosystems
(Mack et al., 2000), and may increase the likelihood of plant invasions. A 2018 student research
project at the University of British Columbia, found that habitats disturbed by humans were more
likely to be dominated by invasive plants than intact forests (Bautista et al., 2018). In a recent
study published in Biological Invasions, Liedtke et al., (2020) looked for invasive species by
distance from trailheads in mountainous areas (Liedtke et al., 2020). They found that trails act as
a conduit for the spread of invasive plants, though the prevalence of invasive plants decreased
with distance from the trailhead (Liedtke et al., 2020). In a different article published in Applied
Vegetation Science, Aziz et al., (2023) also sought to determine if trailheads acted as conduits for
the spread of invasive plant species along the trail (Aziz et al., 2023). They found that invasive
plant cover was highest near trailheads, but decreased with distance from the trailhead; however,
contrary to their predictions, they found that in the forest interior, invasive plant cover increased
with distance from the trailside (Aziz et al., 2023). Moreover, researchers at the University of
British Columbia, in association with Metro Vancouver, studied the presence of invasive species
in three types of greenspaces: parks and recreation areas, natural areas, and leisure facilities
(Nguyen et al., 2021). They found that invasive plant species were associated with median
household income, gardening expenditures, and greenspace type (Nguyen et al., 2021). They also
found more invasive species in areas with higher levels of disturbance, and higher levels of
human activity (Nguyen et al., 2021). These findings are relevant to my study hypothesis that
outdoor recreation areas with higher levels of human activity, should have more invasive plants
present.
North Vancouver has a large number of recreational areas, giving many opportunities for
invasive plants to spread. Although inventories of invasive species in North Vancouver have
been conducted (District of North Vancouver, 2024), no studies have examined the association
between invasive species and use of and proximity to recreational trails. My study focused on six
invasive species known to be present in North Vancouver: Hedera helix (English Ivy), Ilex
aquifolium (English Holly), Prunus laurocerasus (English Laurel), Rubus armeniacus
2

(Himalayan Blackberry), Daphne laureola (Spurge Laurel), and Vinca minor (Periwinkle). In
addition to these six invasive species, I counted any others observed in the study area, (as noted
by the Invasive Species Council of British Columbia). Similar to the Aziz and colleagues (2023)
study, I will measure 3 transects per trail at ~100m intervals with 100m transects perpendicular
to the trailside. Using the methods described by Liedtke and colleagues (2020) to determine the
percent cover of the forest, I estimated the relative amount of light in the study area as it may
play a role in the density of invasive species at a given location. The data was not collected at the
same time of day for all quadrats, thus the amount of light would vary at each location. Relevant
to my study, are how invasive species reproduce, grow, and how they may be found in the study
area. Table 1 provides important information about six of the most common invasive plants in
North Vancouver.

Table 1: Invasive Species and growth information
Species

Scientific Name

Dispersal

Growth

Habitat

Method
English Ivy

Hedera helix

Wind,

Fruits

Soil Type

(Yes/No)
Vegetative

Animals

Forest edge/

Yes (after

Well

fragments

maturity)

drained,
loamy

English

Ilex aquifolium

Holly
Spurge

Wind, Birds,

Vegetative

Mammals
Daphne laureola

Laurel

Forest – edge and
interior, garden

Wind, Birds,

Single stem, or

Garden, forest

Mammals

multi branching

understories

Monopodial

Garden, forest,

English

Prunus

Birds, small

Laurel

laurocerasus

Mammals

Himalayan

Rubus

Birds,

Blackberry

armeniacus

Mammals

Yes

Well
drained

Yes

Various

Yes

Various

Yes

Well

edge
Vegetative

Forest edge, road
side, anywhere

drained

with full sunlight
Periwinkle

Vinca minor

Water,

Vegetative

Wind, Ants

Forest

No (not

Well

edge/fragment

frequently)

drained

See Figure 1 to see some common invasive species in North Vancouver.

3

Figure 1: Images of some common invasive plants. Left: Spurge Laurel, Middle: English Holly and Ivy,
Right: Himalayan Blackberry and Periwinkle.

My study investigated whether the density of invasive species varied with distance from trails
categorized by high-use and low-use intensity. I hypothesized that human activity on outdoor
recreation trails are responsible for the introduction and dispersal of invasive plants into the
forest. I predicted that: 1) high usage intensity trails will have more invasive species present off
trail than low usage intensity trails, 2) deeper into the forest interior and further away from the
trail, I expected to see fewer invasive species with greater distance, and 3) I anticipated that areas
with higher forest cover would have fewer invasive species. Of the five invasive species I was
looking for, I expected that English holly (Ilex aquifolium) would be the most represented out of
all of the invasive species present in North Vancouver. The outcomes of my study may support
the idea that trail use allows the spread of invasive plants and further may inform future
management strategies of areas that are the most affected by invasive species.

Methods
Site Selection
Choosing locations for data collection started with identifying high-usage and low-usage
intensity trails. To determine these trails, I used Strava, which is a social media fitness app where
individuals can track their activities. On Strava, there is a global heat map which provides users
options to show how much each route is used. I set my heat map to red for high-usage, and light
blue for low-usage intensity. After identifying suitable high and low trails I marked their location
on Gaia GPS. Using this program, I assessed trail choices using three criteria. First, if the trail
was at least 500m long so that I could perform three transects on it. Second, I looked to see if
there were any physical barriers such as slopes or riparian areas. Finally, I selected trails that
were at least 500m away from other sampling locations to make sure that two trails were
4

independent from each other. Of 20 potential sites, I ultimately included 12 sites based on these
criteria. (See Figure 2 for Site Selection images)

Figure 2: Site selection method. Left: Strava global heatmap where red is high usage and blue is low
usage. Right: Gaia GPS showing sampling locations and study area.

Data Collection
After marking out prospective sampling locations based on the criteria listed above, data
collection began. At each sampling location, I collected data along three transects spaced at least
100m apart. Each transect was 100m long, starting at the trail and heading perpendicularly into
the forest, as estimated on the Gaia GPS app. On each transect, I collected data within two 20m2
quadrats, one close to the trail (10-30m from trail) and one far from the trail (70-90m from trail).
In each quadrat I counted either the number of stems or estimated percent cover of invasive
plants. The rationale for sometimes counting stems and estimating percent cover was to account
for the differences in the structure of invasive plants found. See Table 2 for a breakdown of how
I collected data for each species found. In each quadrat, I calculated the density value for each
invasive plant found by dividing the number of stems (or percent cover) by the total area of
400m. In addition, in each quadrat I calculated the combined density as the sum of all species
densities. By performing two quadrats, I aimed to capture whether there was variation in two
areas at varying distances from the trail. In addition, at each quadrat I estimated the forest cover
as a percentage out of 100%. In doing this, I would try and capture the relative level of light in
the forest.

Table 2: Data collection method for individual species
Species
English Holly
English Ivy
English Laurel

Type of Measurement
Counting stems/# of plants
Percent cover estimate
Counting stems/# of plants

5

Structure of Plant
Shrub/tree
Vine/shrub
Shrub/small tree

Wall Lettuce

Percent cover estimate

Weed

Data Analysis
To test whether the density of invasive species differed significantly between high and low use
trails, I conducted an ANOVA, with the response variable as density of invasive plants. In this
analysis, I combined the density of the two quadrats for each transect, so each transect had a
single density value. With the data, I conducted two different analyses, one with the combined
density as the response variable, and a second with only the density of holly as the response
variable. English holly was the most prevalent species found in my data collection. The
independent variable for each analysis was trail usage (high or low). Analyses were performed
using Excel, and visualized by box plots.
In the second part of the data analysis I performed two tests using two-way repeated measures
ANOVA with combined density and density of holly set as the response variables, and intensity
and distance as the independent variables (factors). Before conducting these analyses, I tested the
assumptions of normality using the Shapiro Wilk test. The results from the Shapiro Wilk test
showed that the data for both combined density and density of holly were not normally
distributed. To account for this, I transformed both data sets by using a square root
transformation. After the transformation, I performed two more Shapiro Wilk tests, and while the
transformations did not make either data set normally distributed, they were more normally
distributed than before. The data was then visualized using boxplots. All analyses were
conducted using R v 4.4.3 (R Core Team, 2023). See Appendix 1 for full code.
In the final part of data analysis, I performed a correlation analysis with density of invasive
plants as the response variable, and % forest cover as the independent variable. I did two
correlation analyses, one with the combined density, and the other with density of holly. The two
analyses used both high and low usage trails. This analysis took place on Excel, and yielded
scatter plots. In addition to visualizing the data, I found the Pearson correlation coefficient (r)
using the function ‘=CORREL’. On the correlation graphs, I included a trendline to show what
kind of correlation might be present.

Results
Results from the ANOVA analysis indicated that there was no significant difference between
high and low-usage trails. The results were not significant when using data combined across all
species, nor was it significant for just the density of holly. With a level of significance (a) 0.05,
the p-value for the combined density test was 0.527, which is much greater than 0.05. At the
same level of significance, the p-value for the holly test was 0.623 which is also much greater
than alpha. Figures 3 and 4 visualize the results from the two analyses.

6

Figure 3: Boxplot showing no significant difference in densities of high and low usage trails.

7

Figure 4: Boxplot showing no significant differences in density of Holly on high and low usage trails.

Results from the second analysis revealed that there was no significant difference between any of
the factors when using combined density data. The two-way repeated measures ANOVA yielded
results for the two independent variables intensity and distance, and the interaction between the
two. At a level of significance of 0.05, the p-values were 0.962, 0.323, and 0.094 for intensity,
distance, and their interaction, respectively. The results for this can be seen in Figure 5. For the
same analysis with density of holly, there was also no significant difference between any of the
factors. The p-values were 0.879, 0.521, and 0.073 for intensity, distance, and their interaction,
respectively. The results for this part can be seen in Figure 6.

8

Figure 5: Boxplot looking at dependent variable combined density, and independent variables intensity
and distance. H-N = high near, H-F = high far, L-N = low near, L-F = low far. The box plot shows no
significant difference in any of the groups. I will note that there is an interesting interaction going on
between L-N and L-F where the median combined density is larger for L-N than H-N. Further, the
median density is larger in H-F compared to H-N.

9

Figure 6: Boxplot looking at dependent variable density of holly, and independent variables intensity and
distance. H-N = high near, H-F = high far, L-N = low near, L-F = low far. There is an interesting trend
where L-N have a greater median density than H-N. Another interesting trend is that the median density
is larger in H-F, compared to H-N.

The final part of the data analysis looked for correlation between percentage of forest cover and
density of all species combined, and density of holly on both high and low usage trails. On high
usage trails, for both categories, forest cover was negatively correlated with density (see Figure
7). For density of holly alone, r = -0.3248, and for combined density, r = -0.3259. On low usage
trails, forest cover was also negatively correlated with combined density, and density of holly
(see Figure 8). For density of holly, r = -0.1479, and for combined density, r = -0.1542. In both
high and low usage trails, the r is close to 0, indicating weak correlation. By only a slight margin,
trails that are highly used have a stronger negative correlation with density of invasive species.

10

Figure 7: Scatter plot showing the negative correlation for a) Density of Holly and b) Density of all
species combined on High Usage Trails

Figure 8: Scatter plot showing the negative correlation between % Forest Cover and a) density of Holly
and b) density combined across all species on Low Usage Trails

Discussion
Results from the first data analysis found no significant difference in density of invasive plants
between trails categorized by high-usage and low-usage. This was consistent for combined
density of all species identified, and density of holly alone. The two non-significant results
indicate that the hypothesized interaction is not supported, however, there may be reasoning to
explain this. I predicted that trails highly used by outdoor recreation activities would have more
invasive species present, and in general, I noticed more invasive plants. However, not all were
located within the quadrat I used for observation. On low-use trails, I observed invasive plants
present on the trail, but often they did not proliferate into the forest. Another observation while
doing the transects was that if there was a slope, either up or downhill, there were fewer invasive
species present on the slope. Often, I would find a patch of plants at the either the top or bottom
of the slope but none as I went along. This trend was consistent during the data collection. In
11

addition, I observed that where there was a stream or small flowing water body, invasive plants
typically were growing there. My observation is consistent with research by Aronson et al.,
(2017) who found in their study that 55% of riverbank vegetation in the study areas were made
up of non-native (invasive) plants.
The results from the second data analysis reveal that there is not a significant statistical
difference between any of the four groups: high-far, high-near, low-far, and low-near. Aside
from this, there are still important trends to consider. In Figures 5 and 6, the median for both
combined and holly densities are higher in quadrats on low vs high usage trails. This shows that
there is not a consistent trend delineating high and low usage trails with respect to density of
invasive plants. Contradictory to my prediction that there would be more invasive species closer
to, rather than farther from the trail, far quadrats on high usage trails had a higher median density
than near quadrats. This may indicate that the invasive plants I observed can establish better in
forest areas with mild disturbance. Human activity may still play a role in the dispersal of
invasive plants off trail, however, in future research I would consider other variables such as
dispersal mechanisms that may play a greater role. Other studies have found that trails may act as
a conduit for the spread of non-native plants (Liedtke et al., 2020; Aziz et al., 2023), and in my
study, I observed this at varying levels. Visually, there were greater densities of invasive plants
close to the trail, but I did not observe a consistent pattern. For example, at some sites, there were
none at first, but more at second quadrat. At other sites, there was an infestation of invasive
plants at the first, and significantly fewer at the second. When looking at the data before
analyzing it, I saw that in most cases the densities were slightly larger in near quadrats compared
to far, but the difference was not statistically significant. To account for this variation, and small
differences between near and far, the sample size would have to be increased and more trails be
sampled.
In the study, and consistent with my expectations, the data showed that English holly (Ilex
aquifolium) was the most represented invasive species in the forest. Throughout the data
collection process, I noticed that there was plenty of holly scattered along the sides of the trails.
This finding was consistent with Church (2016), that holly is capable of invading forests and
becoming widespread in managed forests. Holly is a shade tolerant evergreen species, but can
also thrive as light becomes available (Church, 2016). Along the trails, I estimated varied levels
of forest cover (level of light), yet I did not find a significant correlation between forest cover
and density of holly. In fact, there was a weak negative correlation between percentage of forest
cover and density of holly. This result is consistent with my prediction that higher levels of forest
cover would see fewer invasive species present. The weak negative correlation suggests that the
level of light is not an important variable in determining density of holly. However, a more
accurate forest cover estimation could be used, which may alter this result. In addition, I looked
at percentage forest cover and combined densities across all species, and there was an even
weaker negative correlation. In association with my predictions, these results make sense

12

because all of the species that I found during data collection are shade tolerant, and can thrive in
the understory. Another important factor to note is the dispersal mechanism of English holly is
primarily by birds, with up to 96% of seeds being dispersed by American Robins (Turdus
migratorius) (Zika, 2010). With this factored in, there should be no pattern in the distribution of
holly in the forest. My results support this, as there is not a significant difference between high
and low usage trails, meaning humans are not likely taking part in the spread of holly throughout
the forest. Moreover, during the data collection process I noted that there was lots of holly
present outside of the quadrats where data was collected. To more accurately determine how
much holly is present in the forest, I could perform a continuous transect that continues further
into the forest and count the number of hollies on both sides. Further, the method of counting
hollies could be refined, as it was difficult to fully capture all of them when some are big trees,
and others are small shrubs.
Throughout my data collection, I only found four invasive species: English holly (Ilex
aquifolium), English ivy (Hedera helix), English laurel (Prunus laurocerasus), and wall lettuce
(Lactuca muralis) at one site. In Table 1, I note several invasive species that have naturalized in
North Vancouver. Two species that I thought would be present, Himalayan blackberry (Rubus
armeniacus) and periwinkle (Vinca minor), were not present in any of the transects sampled.
Himalayan blackberry prefers habitats with well-drained moist soils, often in disturbed sites, and
areas with full sunlight (Gaire et al., 2015). Within the trails sampled, these conditions were not
met, which provided some reasoning to why they were not present. Periwinkle is a forest edge
species, that also prefers moist conditions (Panasenko et al., 2018). The forest that was sampled
was not forest edge, thus why periwinkle was not found in my surveys. As mentioned previously,
percentage forest cover negatively correlated with density combined, and of holly. The measures
of forest cover were estimates, and their accuracy should be considered. However, the negative
correlation indicates that as forest cover increases, the density of invasive species decreases
slightly. This could show that decreasing the light availability makes it harder for some invasive
species to establish (i.e. Himalayan blackberry). However, researchers have found that some
invasive plants can colonize closed-canopy forests at the same level as native species
(Yamamoto & Jones, 2024). This indicates that the presence of invasive species depends on the
site conditions, and if a species can tolerate shady conditions, it will be able to thrive in the
forest.
My initial hypothesis was that human activity and outdoor recreation are responsible for the
accidental introduction and dispersal of invasive plants into the forest. It has been found that
areas with high levels of outdoor recreation and tourism have higher abundance and richness of
invasive species (Anderson et al., 2015). In addition, it has been found that introduced plants can
escape from cultivation and horticulture, and become invasive in the new environment (Reichard
and White, 2001). Thus, these two factors provide reasoning for how invasive plants may have
come to be present in the forest. An outcome of this research is that the relationship between

13

invasive plant species and proximity to recreational trails were not statistically significant.
However, humans have played a role in introducing invasive species and there is evidence of this
along most of the sampled trails. This study is important because it demonstrates that there are
many invasive plants present in the forest which are taking over space and resources from native
plants. Invasive plants are a threat to North Vancouver’s forests; thus, my study provides data
that could inform forest management strategies and public awareness. The outcome of this
research provokes me to think further about why the density of invasive species is higher in
certain areas. Future research could look at how the density of invasive species changes based on
varied site conditions. This could be accomplished by manipulating site conditions over a set
time span. The sites could be disturbed to various levels; no disturbance (control), minor
disturbance (some trampling, small removal of vegetation), intermediate disturbance (trampling,
removal of vegetation), and high disturbance (clear cut, or full removal of vegetation and
terraforming). This could be repeated across several areas, taking into consideration site history,
and other landscape features such as proximity to a riparian area. In addition, it would be
interesting to do an enclosure experiment along trails of varying usage intensities. This
experiment could help determine if trampling from humans, dogs, and other mammals has an
impact on vegetation, and further, if it provides good habitat for invasive species colonization.

Conclusion
Human use of outdoor recreation trails appear not to be responsible for the spread of invasive
species into the forests of North Vancouver. Humans alter the natural landscape, and have a
substantial role in creating conditions for invasive species to thrive. In the case of outdoor
recreation, trails are created by removing vegetation, and transforming the landscape to make it
accessible for hikers and cyclists. This transformation creates optimal habitat, as can be seen in
North Vancouver, where there are many invasive species found along the sides of trails. Further,
this establishment of invasive plants on trails allows them to disperse naturally, and take over
more areas in the forest. Despite finding no distinct pattern in this study, there is something
going on in the forest. Thus, more research must be done to get to the bottom of this issue that
continues to affect native biodiversity in North Vancouver’s forests.

Acknowledgements
Thank you to all of my professors who helped along the way; Caroline Dingle, Mark Vaughn,
Mahta Khosravi and Paul McMillan. A big thank you to Tom Flower, who was instrumental in
helping with the R code for the second analysis. Without Tom, I would not have been able to
finish my data analysis. To my classmates, thank you for the unwavering support and helping me
through the challenges of this project. Finally, I would not be here without the support and
opinions of my friends and family. I thank everyone that helped me along the way, and I am
exceedingly grateful to Capilano University for providing the knowledge that helped me to
complete this project.
14

References
Abrahamczyk, S., Otto, J., & Maximilian Weigend. (2024). The reproductive biology of the neophyte
Prunus laurocerasus in Central Europe. Plant Species Biology. https://doi.org/10.1111/14421984.12500
Anderson, L. G., Rocliffe, S., Haddaway, N. R., & Dunn, A. M. (2015). The Role of Tourism and
Recreation in the Spread of Non-Native Species: A Systematic Review and Meta-Analysis.
PLOS ONE, 10(10), e0140833. https://doi.org/10.1371/journal.pone.0140833
Aronson, M. F. J., Patel, M. V., O’Neill, K. M., & Ehrenfeld, J. G. (2017). Urban riparian systems
function as corridors for both native and invasive plant species. Biological Invasions, 19(12),
3645–3657. https://doi.org/10.1007/s10530-017-1583-1
Aziz, N., Holmes, M. A., Bennion, L. D., Reeb, R. A., & Kuebbing, S. E. (2023). Hiking trailheads,
but not trailsides, associated with higher cover of non‐native plants’ trail impact on non‐native
plant cover. Applied Vegetation Science, 26(1), e12716. https://doi.org/10.1111/avsc.12716
Barney, J. N., Tekiela, D. R., Dollete, E. S., & Tomasek, B. J. (2013). What is the “real” impact of
invasive plant species? Frontiers in Ecology and the Environment, 11(6), 322–329.
https://doi.org/10.1890/120120
Bautista, B., Hu, H., Kaye, M., & Lu, Z. (2018). Comparing Percent Cover of Select Invasive Plant
Species (Rubus armeniacus, Ilex aquifolum, and Hedera helix) at Salish Creek and Canyon
Creek in Pacific Spirit Park. The Expedition, 8.
https://ojs.library.ubc.ca/index.php/expedition/article/view/191467
Herbert, J. (2021). Comparison of English holly (Ilex aquifolium) management methods in Redwood
Park, Surrey, British Columbia Prepared For Rob Landucci NRP Urban Forestry Manager City
15

of Surrey Prepared By.
https://iscmv.ca/docs/Comparison_of_English_holly_management_methods_in_Redwood_Park_
-_FINAL_REPORT.pdf
Christen, D. C., & Matlack, G. R. (2009). The habitat and conduit functions of roads in the spread of
three invasive plant species. Biological Invasions, 11(2), 453–465.
https://doi.org/10.1007/s10530-008-9262-x
Church, E. (2016, April 6). Invasive English holly (Ilex aquifolium L.) in Clear-Cut and Forest Units
in a Western Washington Managed Forest. Washington.edu.
https://digital.lib.washington.edu/researchworks/items/95db109c-afc3-4e1f-a44b-1b5f56791bd1
Coutts, S. R., Van Klinken, R. D., Yokomizo, H., & Buckley, Y. M. (2011). What are the key drivers
of spread in invasive plants: Dispersal, demography or landscape: and how can we use this
knowledge to aid management? Biological Invasions, 13(7), 1649–1661.
https://doi.org/10.1007/s10530-010-9922-5
Davies, K. W. (2011). Plant community diversity and native plant abundance decline with increasing
abundance of an exotic annual grass. Oecologia, 167(2), 481–491.
https://doi.org/10.1007/s00442-011-1992-2
Davies, K. W., & Sheley, R. L. (2007). A conceptual framework for preventing the spatial dispersal of
invasive plants. Weed Science, 55(2), 178–184. https://doi.org/10.1614/WS-06-161
Dennehy, C., Alverson, E. R., Anderson, H. E., Clements, D. R., Gilbert, R., & Kaye, T. N. (2011).
Management strategies for invasive plants in pacific northwest prairies, savannas, and oak
woodlands. Northwest Science, 85(2), 329–351. https://doi.org/10.3955/046.085.0219
District of North Vancouver. (2024). District of North Vancouver 2023 Invasive Plant Inventory (pp.
1–38) [Draft #1].

16

Gaire, R., Astley, C., Upadhyaya, M. K., Clements, D. R., & Bargen, M. (2015). The biology of
canadian weeds. 154. Himalayan blackberry. Canadian Journal of Plant Science, 95(3), 557–
570. https://doi.org/10.4141/cjps-2014-402
Invasive species council of british columbia. (2025, January 23). Invasive Species Council of British
Columbia. https://bcinvasives.ca/
Jiménez-Hernández, H. A., Bonilla-Valencia, L., Martínez-Orea, Y., Zamora-Almazan, M. J.,
Espinosa-García, F. J., & Castillo-Argüero, S. (2023). Effects of Hedera helix L. removal on the
understory early regeneration in an oak temperate forest in Mexico City. Ecological Processes,
12(1), 31. https://doi.org/10.1186/s13717-023-00443-y
Lei, T. (2014). Environmental preferences and constraints of Daphne laureola , an invasive shrub in
western Canada. Canadian Journal of Forest Research, 44(11), 1462–1467.
https://doi.org/10.1139/cjfr-2014-0261
Liedtke, R., Barros, A., Essl, F., Lembrechts, J. J., Wedegärtner, R. E. M., Pauchard, A., & Dullinger,
S. (2020). Hiking trails as conduits for the spread of non-native species in mountain areas.
Biological Invasions, 22(3), 1121–1134. https://doi.org/10.1007/s10530-019-02165-9
MACK, R. N., & LONSDALE, W. M. (2001). Humans as Global Plant Dispersers: Getting More
Than We Bargained For. BioScience, 51(2), 95. https://doi.org/10.1641/00063568(2001)051[0095:hagpdg]2.0.co;2
Mack, R. N., Simberloff, D., Mark Lonsdale, W., Evans, H., Clout, M., & Bazzaz, F. A. (2000).
Biotic invasions: Causes, epidemiology, global consequences, and control. Ecological
Applications, 10(3), 689–710. https://doi.org/10.1890/10510761(2000)010[0689:BICEGC]2.0.CO;2

17

Panasenko, N. N., & Anishchenko, L. N. (2018). Influence of invasive plants parthenocissus vitacea
and vinca minor on biodiversity indices of forest communities. Contemporary Problems of
Ecology, 11(6), 614–623. https://doi.org/10.1134/S1995425518060070
R Core Team. (2023). R: A Language and Environment for Statistical Computing. R Foundation for
Statistical Computing. https://www.r-project.org/
Reichard, S. H., & White, P. (2001). Horticulture as a Pathway of Invasive Plant Introductions in the
United States. BioScience, 51(2), 103. https://doi.org/10.1641/00063568(2001)051[0103:haapoi]2.0.co;2
Sakai, A. K., Allendorf, F. W., Holt, J. S., Lodge, D. M., Molofsky, J., With, K. A., Baughman, S.,
Cabin, R. J., Cohen, J. E., Ellstrand, N. C., McCauley, D. E., O’Neil, P., Parker, I. M.,
Thompson, J. N., & Weller, S. G. (2001). The population biology of invasive species. Annual
Review of Ecology and Systematics, 32, 305–332. https://www.jstor.org/stable/2678643
Schmitz, D. C., & Simberloff, D. (1997). Biological invasions: A growing threat. Issues in Science &
Technology, 13(4), 33.
Wells, F. H., Lauenroth, W. K., & Bradford, J. B. (2012). Recreational trails as corridors for alien
plants in the rocky mountains, usa. Western North American Naturalist, 72(4), 507–533.
https://www.jstor.org/stable/42003924
Yamamoto, M. H., & Jones, C. C. (2024). Comparing long-term patterns of spread of native and
invasive plants in a successional forest. Oecologia, 205(1), 13–25.
https://doi.org/10.1007/s00442-024-05554-7
Zika, P. F. (2010). Invasive hollies (Ilex, aquifoliaceae) and their dispersers in the pacific northwest.
Madroño, 57(1), 1–10. https://doi.org/10.3120/0024-9637-57.1.1

18

Zimmermann, H., Brandt, P., Fischer, J., Welk, E., & Von Wehrden, H. (2014). The Human Release
Hypothesis for biological invasions: Human activity as a determinant of the abundance of
invasive plant species. F1000Research, 3, 109. https://doi.org/10.12688/f1000research.3740.1

Appendix 1
Full R code for second data analysis:
#First set the working directory, there are better ways of importing the data, but this way I can
modify the csv file and then reload to get it into R
#changed back slash to forward slash to fix error
setwd("C:/Users/davidlisle/Documents")
#tell R which is the CSV file to use
DLdata<-read.csv("Betweensubjectsdesign.csv")
library(rstatix)
library(tidyverse)
#briefly look at the data
head(DLdata)
hist(DLdata$Combined.Density)
#transforming the data to seek normality?
DLdata$sqrtCden<-sqrt(DLdata$Combined.Density)
hist(DLdata$sqrtCden)
hist(DLdata$Holly.Density) #visualizing holly density in a histogram
DLdata$sqrtHden<-sqrt(DLdata$Holly.Density) #transforming holly density by using sqrt
function to seek normality
hist(DLdata$sqrtHden) #visualizing transformed data

19

#normality tests - Shapiro wilk
shapiro.test(DLdata$Combined.Density)
shapiro.test(DLdata$sqrtCden)
shapiro.test(DLdata$sqrtHden) #shapiro wilk test for normality for transformed holly density
shapiro.test(DLdata$Holly.Density) #shapiro wilk test for normality for holly density
#code for two-way repeated measures Anova
library(tidyverse)
library(rstatix)
# 1. Convert to factors
DLdata <- DLdata %>%
mutate(
Trail = as.factor(Trail),
Distance = as.factor(Distance),
Intensity = as.factor(Intensity)
)
# 2. Aggregate repeated measures per condition
DL_avg <- DLdata %>%
group_by(Trail, Intensity, Distance) %>%
summarise(
sqrtCden = mean(sqrtCden),
.groups = "drop"
)
# 3. Run the mixed (repeated measures) ANOVA
res.aov <- anova_test(
data = DL_avg,
dv = sqrtCden,
wid = Trail,
within = Distance,
between = Intensity
)
# 4. Show results
get_anova_table(res.aov)
#***Same code for Holly density instead of combined density***
library(tidyverse)

20

library(rstatix)
# 1. Convert to factors
DLdata <- DLdata %>%
mutate(
Trail = as.factor(Trail),
Distance = as.factor(Distance),
Intensity = as.factor(Intensity)
)
# 2. Aggregate repeated measures per condition
DL_avg <- DLdata %>%
group_by(Trail, Intensity, Distance) %>%
summarise(
sqrtHden = mean(sqrtHden),
.groups = "drop"
)
# 3. Run the mixed (repeated measures) ANOVA
res.aov <- anova_test(
data = DL_avg,
dv = sqrtHden,
wid = Trail,
within = Distance,
between = Intensity
)
# 4. Show results
get_anova_table(res.aov)
#Visualizing the two-way repeated measures Anova for Combined.Density and Holly.Density
#setting factors
DLdata$Intensity <- as.factor(DLdata$Intensity)
DLdata$Distance <- as.factor(DLdata$Distance)
DLdata$Trail <- as.factor(DLdata$Trail)
#Compute summary statistics by groups:
#First summary statistics for Combined.Density
DLdata %>% group_by(Distance) %>% get_summary_stats(Combined.Density, type =
"common")
DLdata %>% group_by(Intensity) %>% get_summary_stats(Combined.Density, type =
"common")

21

#Second summary statistics for Holly.Density
DLdata %>% group_by(Distance) %>% get_summary_stats(Holly.Density, type = "common")
DLdata %>% group_by(Intensity) %>% get_summary_stats(Holly.Density, type = "common")
#Visualizing the two-way repeated measure ANOVA for Combined.Density first, then
Holly.Density second. Same code for both
# Create a combined group from Intensity and Distance
DLdata <- DLdata %>%
mutate(Group = interaction(Intensity, Distance, sep = "-"),
Group = factor(Group, levels = c("H-N", "H-F", "L-N", "L-F")))
# Boxplot using Combined.Density, grouped by Intensity-Distance combo
ggplot(DLdata, aes(x = Group, y = Combined.Density, fill = Intensity)) +
geom_boxplot() +
labs(title = "Combined Density by Intensity and Distance",
x = "Intensity + Distance ",
y = "Combined Density") +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
# Create a combined group from Intensity and Distance
DLdata <- DLdata %>%
mutate(Group = interaction(Intensity, Distance, sep = "-"),
Group = factor(Group, levels = c("H-N", "H-F", "L-N", "L-F")))
# Boxplot using Holly.Density, grouped by Intensity-Distance combo
ggplot(DLdata, aes(x = Group, y = Holly.Density, fill = Intensity)) +
geom_boxplot() +
labs(title = "Combined Density by Intensity and Distance",
x = "Intensity + Distance",
y = " Density of Holly ") +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))

22