sediment flux and compaction trends on off

SEDIMENT FLUX AND COMPACTION TRENDS ON OFF-ROAD VEHICLE
(ORV) AND OTHER TRAILS IN AN APPALACHIAN FOREST SETTING
Dorothy Sack
Department of Geography
122 Clippinger Labs
Ohio University
Athens, Ohio 45701
Silvino da Luz, Jr.
Jet Programme
Green Heights Kuratayama 2-2
35-1 Kusube-cho
Ise-shi, Mie-ken 516-001
Japan
Abstract: This research investigates the impact of off-road vehicles (ORVs) on sediment
dynamics and compaction in the forested hillslope terrain of southeastern Ohio. A dense
network of ground-height change, penetrometer resistance, and penetrometer depth measurements was collected from trails and trail-adjacent forest land at four ORV and, for
comparison, two non-ORV trail study sites (hiking, horse-riding). Repeat measurements
were made at the height of the ORV season, at the end of the ORV season, and near the
end of the off-season. Separate statistical comparisons of ground-height change and penetrometer observations for forest versus trail locations, ORV versus non-ORV study sites,
and one time period versus another reveal compaction on all trails, significant surface
compaction on the ORV trails, and considerable sediment flux on both the trails and the
adjacent forest land at the ORV sites. This flux includes sediment transfer during the riding
season from the ORV trails to the adjacent forest land by wash and tire throw. Net erosion
that occurred on the ORV trails during the studied part of the riding season was quantitatively offset by off-season aggradation, but the system is primarily one of sediment
through-put. Considering the entire riding season, erosion rates on the ORV trails may be
as high as 0.11 m3/m2/yr. [Key words: off-road vehicles (ORVs, ATVs, OHVs), landscape
disturbance, recreational impacts, national forest, sediment flux, tire throw, erosion rates.]
INTRODUCTION
As recreational land use has increased in recent decades, so too has research
into the impacts of those recreational activities on the landscape (Leung and
Marion, 1996). This field of study, sometimes referred to as recreational ecology
(Coleman, 1981; Cole, 1989), has emphasized the effects of hiking on the vegetation and soil of more-or-less natural parks and reserves (e.g., Dawson et al., 1974;
Bryan, 1977; Coleman, 1981; Bright, 1986; Cole, 1987; Stewart and Cameron,
1992; Sutherland et al., 2001), but various researchers have also studied the
impacts from the recreational use of horses, pack animals, mountain bicycles, and
536
Physical Geography, 2003, 24, 6, pp. 536–554.
Copyright © 2003 by V. H. Winston & Son, Inc. All rights reserved.
APPALACHIAN TRAIL EROSION
537
motorized off-road vehicles (ORVs) (e.g., Dorrance et al., 1975; Weaver and Dale,
1978; Summer, 1980; Whinam et al., 1994; Wilson and Seney, 1994; Cole and
Spildie, 1998). Thus far, in terms of percentage of publications, geographers have
not played a large role in the recreational impact research arena despite the value
of the geographic perspective in that research (Cole, 1989). Geographers have
instead tended to work on the related topic of erosion from nonrecreational pedestrian paths and trails (e.g., Harden, 1992; Vogler and Butler, 1996; Wallin and
Harden, 1996; Ziegler et al., 2000, 2001).
Results of this combined body of previous research on hiking-dominated recreational and nonrecreational trails show that in many environments trampling
reduces vegetation cover, can decrease plant diversity, can lead to species change,
and causes trail compaction, which encourages runoff and therefore trail erosion
(Dale and Weaver, 1974; Cole, 1978; Bright, 1986; Hall and Kuss, 1989; Sun and
Liddle, 1993; Wallin and Harden, 1996). Trail widening has also been documented
(Dale and Weaver, 1974; Summer, 1986). Some studies have monitored trail properties over time (Summer, 1980, 1986; Fish et al., 1981) and some have compared
trail properties with control areas that are far enough removed from the trail paths
to be unaffected by them (Harden, 1992; Stewart and Cameron, 1992; Whinam et
al., 1994). To our knowledge, previous researchers have only collected data from
the trail-adjacent zone to serve as end points for cross-trail transects; data from the
trail-adjacent zone have not previously been collected to make statistical comparisons with measurements made on the trails per se.
Within the recreational impact literature, the decade from 1974 through 1983
stands out for its flurry of publication activity, mostly from biologists and geologists,
specifically regarding the environmental impacts of off-road vehicles (ORVs). The
bulk of that work concerned arid and semi-arid environments (e.g., Nakata et al.,
1976; Snyder et al., 1976; Vollmer et al., 1976; Iverson, 1980; Iverson et al., 1981;
Adams et al., 1982; Webb and Wilshire, 1983), with a secondary cluster of studies
dealing with coastal sites (e.g., Liddle and Greig-Smith, 1975; Godfrey et al., 1978;
Hosier and Eaton, 1980; Gilbertson, 1981; Chipping and McCoy, 1982). A few
papers addressed alpine and needleleaf forest settings (e.g., Weaver and Dale,
1978), but the physical impacts of wheeled ORVs in humid-region deciduous forests were largely ignored.
Since that prolific decade of ORV research, the general character of the ORV literature has shifted to some extent. Environmentalist essays and scientific research
papers on the environmental impacts of ORVs still appear (e.g., Williams, 2000;
Brown and Schoknecht, 2001), but there has been a growing number of publications by the U.S. federal government and an increasing focus on managing ORVuse areas. With much of the sanctioned ORV use in the U.S. occurring on federal
land, these two trends are not unrelated. Most of the governmental publications
consist of U.S. Department of Agriculture (USDA) and U.S. Department of the Interior (USDI) ORV trail maps, ORV-use public information brochures, and management plans for national forest and range land (e.g., USDA, 1987, 1994, 2000; USDI,
1987, 1989). Thus, there has been some increase in the overall number of ORVrelated publications pertaining to forests. There remains, however, a dearth of independent scientific studies specifically regarding the physical environmental impacts
538
SACK AND DA LUZ
Fig. 1. Generalized location of Wayne National Forest in the eastern United States.
of off-road vehicles on mid-latitude forested land. The research reported on here
contributes to this underrepresented part of the ORV literature.
In this paper we use the term ORV exclusively; the same types of vehicles may
also be called all-terrain vehicles (ATVs) or off-highway vehicles (OHVs). As used in
this paper, ORV refers to motorized vehicles with two, three, or four wheels.
STUDY AREA AND PURPOSE
ORV use is popular across the United States. Much ORV use in the eastern part
of the country occurs in national forests, including the accessible Wayne National
Forest in southeastern Ohio (Fig. 1). Wayne National Forest lies within the
APPALACHIAN TRAIL EROSION
539
Fig. 2. The three units of Wayne National Forest and their proximity to regional population centers.
Appalachian Plateau physiographic province of the Appalachian Highlands, but
also within 200 km of several major urban areas, including Cleveland, Columbus,
Cincinnati, Pittsburgh, and Charleston (Fig. 2). ORV use is permitted in two of the
forest’s three geographical units, the Athens and Ironton units, and is allowed
between April 15 and December 15 of each year. Within the 716 km2 Athens unit,
separate trail systems are maintained for ORVs, horse riding, and hiking. The network of ORV trails encompasses approximately 20% of the land area within the
Athens unit (USDA, 1987). While the horse-riding season coincides with the ORVriding season, hiking is permitted all year. Nevertheless, most hiking also occurs
during the warmer spring, summer, and fall months.
Topography within the Athens unit of Wayne National Forest consists of steep,
stream-dissected ridges underlain by Late Pennsylvanian shales, siltstones, sandstones, limestones, and coals of the Allegheny and Conemaugh Groups (Sturgeon,
1958). Local relief averages about 70 m. The climate of the region is humid continental, and marked by cold winters, hot summers, and about 100 cm of average
540
SACK AND DA LUZ
Fig. 3. Location of studied trail sites within the Athens unit of Wayne National Forest.
annual precipitation that is spread throughout the year. Oak and hickory trees dominate the forest, which also contains some pine and hemlock. Soils in the Athens
unit tend to be deeply weathered, medium-textured, and well-drained (Rubel and
Jenny, 1988; Lemaster and Gilmore, 1989).
This research investigates sediment flux and compaction trends on selected ORV
trails and trail-adjacent forest land in the Athens unit of Wayne National Forest.
Trends determined for the ORV trails are compared with data on sediment flux and
compaction collected from segments of horse-riding and hiking trails that lie in
approximately equivalent local physiographic settings. This research was conducted in order to learn more about the physical impact of ORV use within the
broadleaf deciduous forest environment. Such information can enhance the ability
of forest managers to protect and preserve the biological, soil, and water resources
under their administration.
METHODS
This study examined six 60-m long trail segments within the Athens unit of
Wayne National Forest, four from ORV trails and one each from a horse-riding and
hiking trail (Figs. 3, 4, and 5). All of the studied segments occupy comparable hillslope positions. Characteristics of the field sites are listed in Table 1.
Data collection consisted of repeated measuring of detailed topographic profiles
and soil compaction indicators across the width of the trails at ten sites along each
541
APPALACHIAN TRAIL EROSION
Table 1. Field Site Characteristics
Site
Trail
type
Average
trail width
(m)
Trail
slope
(°)
Trail
aspect
(°)
Hillslope
(°)
Hillslope
aspect
(°)
1
ORV
2.73
12
133
11
65
2
ORV
2.45
11
7
10
125
3
ORV
2.61
19
60
8
5
4
ORV
2.55
8
58
8
135
Pebbly mud
5
Horse
0.95
8
300
6
228
Slightly pebbly fine sandy
mud
6
Hiking
0.75
9
85
5
50
Sediment texture
Granular clay
Fine sandy clay
Muddy fine sand
Slightly granular mud
Fig. 4. A portion of studied trail segment 1, an ORV trail. Note the sediment that has been splashed
onto the lower part of the tree trunk at right by tire throw.
studied trail segment. The cross-trail transect sites were positioned every 6 m along
each trail. At the start of the fieldwork, the two end points of each cross-trail transect
were established and monumented with fixed survey markers so that the specific
locations could be reoccupied and measured on multiple occasions. The fixed survey markers typically consisted of 31-cm (12-inch) metal spikes that were driven
into the ground in the forest well beyond the edge of the trail. Detailed topographic
profiles were measured across the width of the trails but extending on each side of
542
SACK AND DA LUZ
Fig. 5. A portion of studied trail segment 5, a horse-riding trail.
the trail an average of 1.5 m into the adjacent forest land to reach the survey end
points. Ground elevation (y) was measured every 30 cm (x) along each cross-trail
profile at the ORV trails. Because the horse-riding and hiking trails are narrower
than the ORV trails (Table 1), elevation observations along those profiles were made
at 15-cm horizontal increments. Each data point was clearly labelled in the field
notes as being derived from the trailed or the forested part of the cross profile.
Except for the extension of the profile well beyond the trail, similar cross-profiling
field techniques have been used in previous trail studies, such as those by Summer
(1980, 1986), Fish et al. (1981), Jubenville and O’Sullivan (1987), Whinam et al.
(1994), and Vogler and Butler (1996).
Observations on maximum penetrometer resistance and a variable that has
received little previous attention in trampling studies, depth of maximum penetrometer resistance, were collected adjacent to the topographic profile data points as indicators of soil compaction (Adams et al., 1982). A soil that is more compacted will
exhibit greater penetrometer resistance, that is, greater strength, and smaller maximum penetrometer depth than a less compacted soil. Penetrometer resistance was
measured on an ordinal scale with a low, medium, and high category that consisted
of 0–138 N/cm2, 138–207 N/cm2, and values greater than 207 N/cm2, respectively.
For all six of the studied trail segments, complete sets of the topographic profile
and soil compaction data were acquired at three different times, referred to as t1, t2,
and t3. The first (t1) set of measurements was collected during the height of the ORV
season in September of 1998. The second (t2) set of measurements was made near
the end of ORV season in fall of 1998. The third (t3) data set was gathered near the
end of the ORV off-season in March of 1999. The topographic profile data were
APPALACHIAN TRAIL EROSION
543
converted for subsequent statistical analyses to the change in ground height (∆h)
experienced at each profile point between t2 and t1, t3 and t2, and t3 and t1. The
amount and direction of ground-height change experienced during the three periods at the various measurement points indicate the aggradational/degradational
impacts of the ORV season (t2-t1), the ORV off-season (t3-t2), and the study interval
(t3-t1). Note that we are using the term degradation in this paper to mean net erosion rather than in the more general way in which it is used in some recreational
trail impact literature (Leung and Marion, 1996). Although an increase in soil compaction might cause a slight decrease in ground height, for the purposes of this
paper observed changes in ground elevation are assumed to be only the result of net
sediment erosion in the case of surface degradation and net sediment deposition in
the case of surface aggradation.
Statistical tests were used to compare the observations on each of the three
variables—profile-point elevation change, maximum penetrometer resistance, and
depth of maximum penetrometer resistance—for (1) trail versus forest sectors of the
profiles, (2) ORV versus non-ORV trail sites, and (3) one time period versus another.
Nonparametric statistical tests were employed because the observations are not
normally distributed. The Mann-Whitney U test was selected for analyzing the
ground-height change and the maximum penetrometer depth data because those
variables consist of interval- or ratio-scale data. The maximum penetrometer resistance data, which were collected at the ordinal scale, meet the requirements of the
Kolmogorov-Smirnov test (Siegel, 1956). All statistical tests were conducted at the
.05 level of significance.
DESCRIPTIVE STATISTICS
Table 2 lists maximum degradation, maximum aggradation, and mean groundheight change values determined separately for the forest and trail sections of each
study site for the three time intervals. It also shows those values for the combined
group of all ORV study sites. These mean and extreme figures help in characterizing
the magnitude and direction of surface elevation changes experienced by the various subsamples. Thus, they also help in generating research hypotheses regarding
differences within and between the distinct trail types. Overall, ground-height
change extends from a maximum degradation value of -37.0 cm, which occurred at
an ORV site trail point, to a maximum aggradation value of 27.5 cm, recorded at an
ORV site forest point. Mean values of ground-height change range from -3.94 cm to
2.82 cm, derived from the trail portion of ORV site 2 and the forest portion of ORV
site 3, respectively. Note that none of the extreme values is associated with the
horse-riding or the hiking site.
On the basis of location and time period, mean ∆h values for the combined data
set of all ORV sites suggest that ORV trail areas experienced net deposition in the
trail-adjacent forest zone especially during the riding season (t2-t1), but also during
the off-season (t3-t2), leading to forest aggradation for the entire study interval (t3-t1).
The all-ORV mean ∆h values for the trail paths point to net trail erosion during the
riding season, followed by net deposition during the off-season. Averages from the
horse-riding site show aggradation in all studied intervals for both the trail-adjacent
544
SACK AND DA LUZ
Table 2. Mean and Extreme Ground-Height Change (∆h) Values for Trail and
Forest Points by Study Site and Time Interval
Trail points
Site
Forest points
Maximum Maximum Mean
degradation aggradation ∆h
(cm)
(cm)
(cm)
n
Maximum Maximum Mean
degradation aggradation ∆h
(cm)
(cm)
(cm)
n
All ORV sites
t2-t1
-23.1
26.5
-0.78
435
-25.7
24.8
0.57
436
t3-t2
-32.2
14.0
1.08
435
-20.4
17.8
0.13
436
t3-t1
-37.0
23.0
0.30
435
-25.7
27.5
0.71
436
t2-t1
-16.3
20.9
-0.25
123
-7.8
5.8
-0.26
108
t3-t2
-26.3
10.5
0.00
123
-9.2
14.8
0.29
108
t3-t1
-14.4
12.0
-0.25
123
-11.4
15.5
0.03
108
t2-t1
-23.1
5.2
-3.94
111
-25.7
3.6
-2.67
110
t3-t2
-32.2
14.0
1.31
111
-5.5
6.2
0.24
110
t3-t1
-37.0
9.4
-2.63
111
-25.7
7.2
-2.43
110
t2-t1
-22.5
26.5
-0.31
103
-10.3
24.8
2.44
100
t3-t2
-14.0
12.3
2.16
103
-20.4
16.9
0.33
100
t3-t1
-15.4
23.0
1.90
103
-14.5
27.5
2.82
100
t2-t1
-6.4
11.4
1.63
98
-7.3
14.0
2.75
118
t3-t2
-5.0
9.0
1.10
98
-10.6
17.8
-0.30
118
t3-t1
-2.5
10.9
2.68
98
-2.3
16.8
2.50
118
t2-t1
-4.7
4.8
0.25
63
-7.7
5.7
0.26
135
t3-t2
-4.8
5.9
0.37
63
-5.0
8.6
0.17
135
t3-t1
-4.9
4.3
0.62
63
-6.0
5.0
0.43
135
t2-t1
-3.8
11.8
0.75
50
-13.2
10.6
-0.19
150
t3-t2
-5.3
7.2
0.76
50
-8.5
17.5
0.87
150
t3-t1
-2.2
6.7
1.51
50
-4.0
6.5
0.69
150
1 ORV
2 ORV
3 ORV
4 ORV
5 Horse
6 Hiking
forest sector and the trail path itself. At the hiking site, mean ∆h values indicate that
aggradation occurred on the trail sector for all studied intervals. The trail-adjacent
forest zone at the hiking site displayed degradation, on average, from t2 to t1 followed by sufficient aggradation in the ORV off-season to result in mean aggradation
for the study period as a whole (t3-t1).
545
APPALACHIAN TRAIL EROSION
Table 3. Mean Values of Maximum Penetrometer Resistance and
Maximum Pentrometer Depth for Trail and Forest Points by Study Site and
Time Interval
Penetrometer resistance (N/cm2)
Penetrometer depth (cm)
Trail
Forest
n
Trail
Forest
n
t1
197
173
435
10.1
23.8
435
t2
201
186
435
13.0
29.7
435
t3
198
186
435
39.2
50.0
435
t1
207
203
108
9.1
18.5
123
t2
207
206
108
16.0
35.1
123
t3
199
180
108
38.1
49.7
123
t1
193
152
110
10.8
14.4
111
t2
206
195
110
16.7
23.7
111
t3
207
206
110
58.0
58.9
111
t1
181
128
100
12.0
28.7
103
t2
182
134
100
10.7
27.3
103
t3
178
149
100
16.2
33.0
103
t1
207
204
118
8.4
32.9
98
t2
207
202
118
7.8
32.3
98
t3
206
205
118
43.9
56.6
98
t1
207
202
135
16.5
21.5
63
t2
207
204
135
17.6
32.4
63
t3
207
207
135
61.1
62.3
63
t1
207
207
150
20.3
27.6
50
t2
207
201
150
20.7
27.1
50
t3
207
202
150
38.3
39.7
50
Site
All ORV sites
1 ORV
2 ORV
3 ORV
4 ORV
5 Horse
6 Hiking
Table 3 provides mean values of the penetrometer variables for the various subsamples. Penetrometer resistance means were calculated using 69 N/cm2 to represent each observation in the low category, 138 N/cm2 for each observation in the
medium category, and 207 N/cm2 for observations in the high resistance category.
The forest zone of ORV site 3 had the lowest mean resistance at each measurement
546
SACK AND DA LUZ
Table 4. Significant Results (α < .05) of One-Tailed Statistical Tests Comparing
Ground-Height Change and Soil Compaction Variables of Trail (T) Versus
Forest (F) Samplesa
Ground-height change
Penetrometer resistance
Maximum penetrometer
depth
Site
t2-t1
t3-t2
t3-t1
t1
t2
t3
t1
t2
t3
All ORV sites
T(-)
T(+)
–
T
T
T
F
F
F
1 ORV
–
–
–
T
–
T
F
F
F
2 ORV
–
T(+)
–
T
–
–
F
F
–
3 ORV
F(+)
T(+)
–
T
T
T
F
F
F
4 ORV
F(+)
T(+)
–
–
–
–
F
F
F
5 Horse
–
–
–
T
T
–
F
F
–
6 Hiking
T(+)
–
T(+)
–
T
T
F
F
–
a
T or F indicates the sample with values of larger magnitude in cases of significant difference; a
dash designates no statistically significant difference. (+) or (-) denotes whether the sample of
greater ground-height change was dominated by aggradation or degradation, respectively. Sample
sizes appear in Table 2.
time while the trailed zone of both the hiking and the horse-riding site yielded the
maximum possible mean resistance for all three measurement times. Means of the
maximum penetrometer depth in the forest zones ranged from 14.4 to 58.9 cm at
the ORV sites, and up to 62.3 cm in the forest at the horse-riding site. On the trails
per se, mean penetrometer depth extended from 7.8 to 61.1 cm, again with the
highest value occurring at the horse-riding site. The penetrometer means presented
in Table 3 imply that trail paths may be more compacted than the trail-adjacent forest zones, that the horse-riding and hiking sites tend to have greater penetrometer
resistances than the ORV sites, but that the maximum compaction values occur farther below the ground surface on the horse-riding and hiking trail paths than they
do on the ORV trail paths.
Overall, the descriptive data presented in Tables 2 and 3 suggest that there may
be some important differences in (1) forest versus trail sections within some of the
study sites, (2) forested portions of ORV versus non-ORV sites, and (3) trailed sectors
of ORV versus non-ORV sites. The means in Tables 2 and 3 were used to formulate
one-tailed research hypotheses that were then tested for statistical significance.
RESULTS OF STATISTICAL ANALYSES
Forest Versus Trail Sectors
Table 4 summarizes the results obtained by statistically comparing groundheight change, penetrometer resistance, and penetrometer depth observations
between trail and forest sectors at each study site by time period. Parallel analyses
were also conducted on the forest versus trail observations from all ORV sites combined. Because trends that exist at the individual ORV sites are incorporated into
APPALACHIAN TRAIL EROSION
547
the all-ORV group, the ORV study site part of this discussion focuses on results
derived from that more general, combined set of all-ORV data.
Compared to the forest land adjacent to the ORV trails, the combined set of allORV data shows that the ORV trail paths experienced statistically significant degradation during the ORV-use season (t2-t1) and aggradation during the ORV offseason (t3-t2). Meanwhile, the forest land adjacent to the ORV trails underwent net
deposition in both seasons, but especially during the time of ORV use, as indicated
by the subsample means (Table 2). ORV use appears to contribute to aggradation on
forest land adjacent to the ORV trails as well as to erosion of the trails. When considering the all-ORV data over the entire study period (t3-t1), which lasted from the
height of the ORV season until almost the end of the off-season, the total net change
in ground elevation did not differ between the forest and trail sectors (Table 4). Net
erosion on the ORV trails during the studied part of the riding season was counterbalanced by the net deposition that occurred there during the vehicle rest period.
Further confirmation of this seasonal change from trail degradation to trail aggradation comes from comparing the all-ORV trail-zone data of t2-t1 with that of t3-t2
using the Mann-Whitney U test. Those results (not tabulated) verify at p < .000001
that the positive ground-height change experienced on the ORV trails during the
off-season is significantly different from the on-season negative elevation change.
Meaningful differences in profile change were not found between the trailed sector of the horse site and the forested zone bordering the horse trail over any of the
studied time intervals. Apparently both environments underwent moderate
amounts of aggradation during and after the riding season (Tables 2 and 4). Since
the forest zone at the hiking and ORV sites also showed off-season aggradation
(Table 2), net deposition is probably the natural, background behavior of the studied
slopes in the off-season. Moreover, because net deposition on the horse trail during
the riding season is statistically similar to the net deposition on the adjacent forest
land at the same time, it cannot be concluded that horse riding is causing trail
aggradation. Instead, erosion from horse riding is insufficient to offset the natural
background level of aggradation.
The Mann-Whitney U test confirms the hypothesis generated from the descriptive statistics that the trailed portion of the hiking site experienced net deposition
between t1 and t2 while at the same time the adjacent forest zone underwent net
erosion (Tables 2 and 4). Because rest-period (t3-t2) aggradation in the forest sector
adjacent to the hiking trail was similar in amount to rest-period aggradation on the
hiking trail, the hiking trail path exhibited statistically significant greater accumulation than the adjacent forest land for the study period as a whole (t3-t1). Fish et al.
(1981, p. 396) likewise found net deposition on hiking trails relative to control sites
in Guadalupe Mountains National Park in Texas, which they postulate might be
caused by trails impeding normal drainage routes.
The forest versus trail statistical analyses reveal a strong tendency for the trail
paths of all three types of recreational use to have greater maximum penetrometer
resistance than the adjacent forest land. Regardless of the type of recreational use,
penetrometer resistance of the forest soils never exceeded that of the trails. Even
more impressive is the almost complete dominance of the forest soils in displaying
maximum penetrometer depth. Results concerning both penetrometer variables
548
SACK AND DA LUZ
Table 5. Significant Results (α < .05) of One-Tailed Statistical Tests Comparing
ORV Trails with Horse Trails and Hiking Trailsa
Non-ORV
site
Ground-height change
Penetrometer resistance
Maximum penetrometer
depth
t2-t1
t3-t2
t3-t1
t1
t2
t3
t1
t2
t3
Trail
–
ORV(+)
–
Horse
Horse
Horse
Horse
Horse
Horse
Forest
–
–
–
Horse
Horse
Horse
–
–
Horse
5 Horse
6 Hiking
Trail
–
–
–
Hiking
Hiking
Hiking
Hiking
Hiking
–
Forest
–
Hiking(+)
–
Hiking
Hiking
Hiking
Hiking
ORV
ORV
a
In cases of statistically significant differences, the denoted sample is the one with larger values. (+)
indicates that the sample of larger ground-height change was dominated by aggradation; a dash
represents no statistically significant difference. Sample sizes appear in Table 2.
clearly point to greater compaction of the trailed compared to the forested land for
all three types of trails (Table 4).
ORV Versus Non-ORV Sites
Statistical comparisons of the group of ORV study sites with the horse-riding site
reveal only limited differences in their ground-height change history but substantial
differences in their penetrometer characteristics (Table 5). Ground-height change in
the forest sector of the horse-riding site displayed similar amounts of net deposition
in all three time periods to what was observed in the forest sector of the all-ORV
data set (Tables 2 and 5). Although the trailed portion of the horse-riding site underwent a net increase in average surface elevation during the riding season (t2-t1)
when the ORV trails recorded net erosion, that difference is not statistically significant at the .05 level. A statistically significant difference in trail height change
between the horse-riding and the ORV trail paths occurred only in the off-season
when the group of ORV trails experienced greater net deposition than the horse trail
(Tables 2 and 5). That there was significantly greater penetrometer resistance on the
forest and trail portions of the horse-riding site than on equivalent parts of the allORV group at the beginning (t1), middle (t2), and end (t3) of the study is confirmed
by the Kolmogorov-Smirnov test. Nevertheless, despite greater soil strength, the
horse-riding site yielded significantly greater maximum penetration depths than the
ORV sites at all measurement times on the trail and at t2 for the forest.
Statistical tests determined that the forested zone at the hiking site experienced
markedly greater net deposition than the forested part of the ORV study sites during
the ORV off-season. Ground-height change on the trailed portion of the hiking site
was not found to differ statistically from ground-height change on the trails at the
ORV study sites on or off season (Table 5) despite their different mean values (Table
2). As was the case with the horse-riding site, the hiking site had greater penetrometer resistance, that is, greater soil strength, than the all-ORV group for both forest
APPALACHIAN TRAIL EROSION
549
and trail sections at all measurement times. Penetrometer depth of the hiking trail
path, however, statistically exceeded that of the trails in the all-ORV group during
the middle (t1) and at the end (t2) of the ORV season. Penetrometer depth in the forest zone adjacent to the hiking trail also exceeded that adjacent to the ORV trails at
the peak of the riding season (t1), but was significantly less than that adjacent to the
ORV trail at the end of the ORV season (t2) and at the end of the off-season (t3).
DISCUSSION
Repeat measuring of detailed topographic profiles across the width of the trails
and into the adjacent forest land reveals considerable sediment flux at the studied
ORV trail sites. Analysis of the all-ORV data set shows that during the riding season
net erosion on the ORV trail paths was accompanied by net deposition on the adjacent forest land. The notion that some of the sediment eroded from the ORV trails
was deposited directly on the adjacent forest land is supported by the amount of
aggradation that occurred in the forest zone during the ORV season and by the fact
that, according to the Mann-Whitney U test, this on-season forest-sector aggradation is significantly greater than off-season forest aggradation (p = .03). It is also corroborated by field observations of sediment being thrown into the air by ORV tires
and by the existence of unvegetated sediment coating understory plants and the
trail-facing sides of forest-zone tree trunks (Fig. 4). In addition to this anthropomechanical sediment-splash process, the compacted state of the ORV trails would
have decreased infiltration and increased surface runoff on the paths resulting in the
fluvial translocation of sediment from the ORV trails to adjacent land downslope
from the trails (Eckert et al., 1979; Iverson, 1980; Harden, 1992; Ziegler et al.,
2000). At some study-site locations the gravity vector would have directed this fluvial load to the forest land along the margin of the ORV trails where it would have
been recorded as part of the ORV-season forest aggradation. At other sites, however, the topography would have directed surface runoff and its load more longitudinally down the axis of the compacted, unvegetated trail to points beyond the
measurement tracts. Mass wasting might also have removed some materials from
the ORV trails, but it did not appear in the field to be an important process.
Results of this research show that the degradation that occurred on the Wayne
National Forest ORV trails during the studied period of vehicle use was quantitatively offset during the ensuing off-season by aggradation on those trails. This
replacement of previously eroded sediment, however, does not signify sediment
return or system recovery to initial conditions. It is true that some of the sediment
that was transported by tire throw from the ORV trails into the adjacent forest during
the ORV season probably washed or gravitated by mass wasting back onto the trails
during the off-season, but this could only happen from those adjacent sites that
slope toward the trail. Because the studied trail segments generally cross the hillslope contours at oblique angles (Table 1), roughly half of the measured forest land
slopes away from the trail. In this hillslope terrain, sediment eroded primarily by
wash and tire throw from the ORV trails during the studied part of the riding season
had to have been replaced in the off-season mainly by net deposition of new sediment transported in from upslope forest and trail areas. Thus, the temporal
550
SACK AND DA LUZ
variations in trail and forest ground height recorded at the ORV sites reveal a system
of sediment through-put for the study period. Presumably, the slope processes that
deliver sediment to the ORV trails in the off-season operate during the riding season
as well. During the riding season, however, sediment delivery to the trails is greatly
exceeded by trail erosion. Had this research captured the entire eight months of the
annual ORV season, instead of only its last three months, it would likely have documented a system of greatly accelerated erosion rather than an interval of sediment
through-put. With the ORV-season data as collected, the largest mean net erosion
(degradation) recorded for an ORV trail is -3.94 cm at ORV site 2 (Table 2). Applying this mean to the entire 2.45-m wide and 60-m long trail segment over the full
riding season yields an estimated 15.44 m3 of sediment eroded from that trail segment during the eight-month ORV season. Assuming no additional erosion
occurred in the off-season, this would represent an annual trail erosion rate of about
0.11 m3/m2. At an average bulk density of 1900 kg/m3, the sediment loss from the
trail under these conditions would be 209 kg/m2/yr.
The trends that appear in the all-ORV data are represented to varying extents in
the results from the individual ORV sites (Table 4). The individual ORV sites closely
follow the all-ORV data set in displaying larger mean penetrometer resistance on
the trail and larger mean penetrometer depth in the forest (Tables 3 and 4). Clearly,
the penetrometer was inserted into the trail-adjacent forest ground more easily than
into the ORV trails. Compared to these results for the compaction indicators, the
individual ORV sites display more variability in the direction and magnitude of
mean ground-height change for the various time intervals (Table 2). This may be
partly due to differences in site characteristics (Table 1) (Summer, 1986). Results of
the statistical tests for the individual sites, however, generally support, and in no
cases contravene, the results obtained from the all-ORV data set.
The horse-riding and hiking sites behaved quite differently than the group of
ORV sites in this study particularly with respect to ground-height change. The nonORV sites did not display the concomitant trail degradation and forest aggradation
that was experienced by the group of ORV sites during the riding season (t2-t1)
(Table 4). Neither did they exhibit the sequential pattern of on-season trail degradation counterbalanced by off-season aggradation that marks the all-ORV group. The
fact that the direct statistical comparisons of parallel environments at the ORV versus the non-ORV sites indicate only two instances of significant difference for
ground-height change (Table 5) probably reflects the relatively small size of the
non-ORV samples and the conservative nature of nonparametric tests (Siegel,
1956). The fact that both the horse and the hiking trail paths experienced net deposition during the season of use (t2-t1) may indicate that, like the ORV trails, these are
systems dominated by sediment through-put.
As expected, the horse-riding and hiking trails, like the ORV trails, showed
greater compaction than their respective forest zones (Sun and Liddle, 1993); however, when comparing parallel environments between the ORV and non-ORV sites
some unexpected outcomes in the compaction variables appear. Weaver and Dale
(1978) found generally greater compaction on horse trails than on motorcycle trails,
with hiking trails exhibiting the least compaction. In this Wayne National Forest
study, penetrometer resistance at both the horse riding and the hiking site exceeded
APPALACHIAN TRAIL EROSION
551
that at the ORV sites for all time periods for the forest as well as the trail sectors
(Table 5). But despite having stronger soils, which indicates greater compaction, the
horse riding and hiking trail paths also displayed deeper penetrability than the ORV
trails did, and this suggests that the horse-riding and hiking trails had less surface
compaction than the ORV trails. These apparently conflicting outcomes may stem
from some fundamental difference in site characteristics between the ORV and
non-ORV sites. The horse-riding and hiking sites, for example, have slightly lower
gradients and are more mud-dominated than three of the four ORV sites (Table 1).
Alternatively, the recreational activities of horse riding and hiking likely influence
the physical characteristics of the substrate in a different way than ORVs do (Liddle
and Greig-Smith, 1975). This research suggests that although repeated horse riding
and hiking increase soil strength, that impact is not conveyed as completely through
the soil column as the impact of ORVs.
CONCLUSIONS
This research demonstrates that the physical impacts of ORV use in a midlatitude, humid-region forest include significant trail compaction that extends close
to the trail surface, and a considerable amount of sediment flux. Part of this flux
consists of the transfer of sediment during the riding season from the ORV trails to
adjacent forest land primarily by fluvial processes and tire throw. Sediment mobility
is also represented by net deposition on the ORV trails during the off-season, which
signifies a system of sediment through-put, and by trail erosion rates estimated to be
as high as 0.11 m3/m2/yr (209 kg/m2/yr). This rate exceeds one and lies close to the
two other previously published annual maximum erosion estimates of 62 kg/m2/yr
(Wilshire et al., 1978), 173 kg/m2/yr (Griggs and Walsh, 1981), and 250 kg/m2/yr
(Stull et al., 1979), which were determined for ORV trails in semi-arid climates.
Thus, although mid-latitude forest environments are not often perceived as being as
sensitive or as fragile as arid environments (Stebbins, 1974), they appear to be just
as susceptible to landscape disturbance by ORVs, even with the existence of
vehicle-rest periods. Moreover, unlike the case in most arid-region hillslope settings, in the hillslope terrain of a humid climate, such as in Wayne National Forest,
downslope transportation often directs the high sediment yields associated with
ORV use toward perennial stream channels. With the recreational use of ORVs in
American mid-latitude forests well established, officially sanctioned, and expanding, additional research concerning their impact on that environment under actual
conditions of use is clearly warranted.
Acknowledgments: We thank the reviewers of this manuscript for their helpful comments and suggestions.
REFERENCES
Adams, J. A., Endo, A. S., Stolzy, L. H., Rowlands, P. G., and Johnson, H. B. (1982)
Controlled experiments on soil compaction produced by off-road vehicles in the
Mojave Desert, California. Journal of Applied Ecology, Vol. 19, 167–175.
552
SACK AND DA LUZ
Bright, J. A. (1986) Hiker impact on herbaceous vegetation along trails in an evergreen woodland of central Texas. Biological Conservation, Vol. 36, 53–69.
Brown, G. and Schoknecht, N. (2001) Off-road vehicles and vegetation patterning
in a degraded desert ecosystem in Kuwait. Journal of Arid Environments, Vol. 49,
413–427.
Bryan, R. B. (1977) Influence of soil properties on degradation of mountain hiking
trails at Grövelsjön. Geografiska Annaler, Vol. 59A, 49–64.
Chipping, D. H. and McCoy, R. (1982) Coastal sand dune complexities: Pismo
Beach and Monterey Bay. California Geology, Vol. 35, 7–12.
Cole, D. N. (1978) Estimating the susceptibility of wildland vegetation to trailside
alteration. Journal of Applied Ecology, Vol. 15, 281–286.
Cole, D. N. (1987) Effects of three seasons of experimental trampling on five montane forest communities and a grassland in western Montana, USA. Biological
Conservation, Vol. 40, 219–244.
Cole, D. N. (1989) Recreational ecology: What we know, what geographers can
contribute. Professional Geographer, Vol. 41, 143–148.
Cole, D. N. and Spildie, D. R. (1998) Hiker, horse and llama trampling effects on
native vegetation in Montana, USA. Journal of Environmental Management, Vol.
53, 61–71.
Coleman, R. (1981) Footpath erosion in the English Lake District. Applied Geography, Vol. 1, 121–131.
Dale, D. and Weaver, T. (1974) Trampling effects on vegetation of the trail corridors
of north Rocky Mountain forests. Journal of Applied Ecology, Vol. 11, 767–772.
Dawson, J. O., Hinz, P. N., and Gordon, J. C. (1974) Hiking-trail impact on Iowa
streamvalley forest preserves. Iowa State Journal of Research, Vol. 48, 329–337.
Dorrance, M. J., Savage, P. J., and Huff, D. E. (1975) Effects of snowmobiles on
white-tailed deer. Journal of Wildlife Management, Vol. 39, 563–569.
Eckert, R. E., Jr., Wood, M. K., Blackburn, W. H., and Peterson, F. F. (1979) Impacts
of off-road vehicles on infiltration and sediment production of two desert soils.
Journal of Range Management, Vol. 32, 394–397.
Fish, E. B., Brothers, G. L., and Lewis, R. B. (1981) Erosional impacts on trails in
Guadalupe Mountains National Park, Texas. Landscape Planning, Vol. 8, 387–
398.
Gilbertson, D. D. (1981) The impact of past and present land use on a major coastal
barrier system. Applied Geography, Vol. 1, 97–119.
Godfrey, P. J., Leatherman, S. P., and Buckley, P. A. (1978) Impact of off-road vehicles on coastal ecosystems. Coastal Zone '78, Vol. 2, 581–600.
Griggs, G. B. and Walsh, B. L. (1981) The impact, control, and mitigation of off-road
vehicle activity in Hungry Valley, California. Environmental Geology, Vol. 3,
229–243.
Hall, C. N. and Kuss, F. R. (1989) Vegetation alteration along trails in Shenandoah
National Park, Virginia. Biological Conservation, Vol. 48, 211–227.
Harden, C. P. (1992) Incorporating roads and footpaths in watershed-scale hydrologic and soil erosion models. Physical Geography, Vol. 13, 368–385.
APPALACHIAN TRAIL EROSION
553
Hosier, P. E. and Eaton, T. E. (1980) The impact of vehicles on dune and grassland
vegetation on a south-eastern North Carolina barrier beach. Journal of Applied
Ecology, Vol. 17, 173–182.
Iverson, R. M. (1980) Processes of accelerated pluvial erosion on desert hillslopes
modified by vehicular traffic. Earth Surface Processes, Vol. 5, 369–388.
Iverson, R. M., Hinckley, B. S., Webb, R. M., and Hallet, B. (1981) Physical effects
of vehicular disturbances on arid landscapes. Science, Vol. 212, 915–917.
Jubenville, A. and O’Sullivan, K. (1987) Relationship of vegetation type and slope
gradient to trail erosion in interior Alaska. Journal of Soil and Water Conservation, Vol. 42, 450–452.
Lemaster, D. D. and Gilmore, G. M. (1989) Soil Survey of Hocking County, Ohio.
Washington, DC: U.S. Department of Agriculture, Soil Conservation Service.
Leung, Y. F. and Marion, J. L. (1996) Trail degradation as influenced by environmental factors: A state-of-the-knowledge review. Journal of Soil and Water Conservation, Vol. 51, 130–136.
Liddle, M. J. and Greig-Smith, P. (1975) A survey of tracks and paths in a sand dune
ecosystem, part I–soils. Journal of Applied Ecology, Vol. 12, 893–908.
Nakata, J. K., Wilshire, H. G., and Barnes, G. G. (1976) Origin of Mojave Desert
dust plumes photographed from space. Geology, Vol. 4, 644–648.
Rubel, N. and Jenny, P. (1988) Soil Survey of Perry County, Ohio. Washington, DC:
U.S. Department of Agriculture, Soil Conservation Service.
Siegel, S. (1956) Nonparametric Statistics for the Behavioral Sciences. New York,
NY: McGraw-Hill.
Snyder, C. T., Frickel, D. G., Hadley, R. F., and Miller, R. F. (1976) Effects of
Off-Road Vehicle Use on the Hydrology and Landscape of Arid Environments in
Central and Southern California. Washington DC: U.S. Geological Survey WaterResources Investigations 76-99.
Stebbins, R. C. (1974) Off-road vehicles and the fragile desert. The American Biology Teacher, Vol. 36, 203-208, 220, 294–304.
Stewart, D. P. C. and Cameron, K. C. (1992) Effect of trampling on the soils of the
St. James Walkway, New Zealand. Soil Use and Management, Vol. 8, 30–36.
Stull, R., Shipley, S., Hovanitz, E., Thompson, S., and Hovanitz, K. (1979) Effects of
off-road vehicles in Ballinger Canyon, California. Geology, Vol. 7, 19–21.
Sturgeon, M. T. (1958) The Geology and Mineral Resources of Athens County,
Ohio. Columbus, OH: Ohio Division of Geological Survey Bulletin 57.
Summer, R. M. (1980) Impact of horse traffic on trails in Rocky Mountain National
Park. Journal of Soil and Water Conservation, Vol. 35, 85–87.
Summer, R. M. (1986) Geomorphic impacts of horse traffic on montane landforms.
Journal of Soil and Water Conservation, Vol. 41, 126–128.
Sun, D. and Liddle, M. J. (1993) A survey of trampling effects on vegetation and soil
in eight tropical and subtropical sites. Environmental Management, Vol. 17,
497–510.
Sutherland, R. A., Bussen, J. O., Plondke, D. L., Evans, B. M., and Ziegler, A. D.
(2001) Hydrophysical degradation associated with hiking-trail use: A case study
of Hawai’iloa Ridge trail, O’ahu, Hawai’i. Land Degradation and Development,
Vol. 12, 71–86.
554
SACK AND DA LUZ
U.S. Department of Agriculture, Forest Service. (1987) Land and Resource Management Plan, Wayne National Forest. Washington, DC: Government Printing
Office.
U.S. Department of Agriculture, Forest Service. (1994) Hanging Rock and Pine Creek
Off-Road Vehicle Guide: Ironton District, Wayne National Forest. Washington,
DC: Government Printing Office.
U.S. Department of Agriculture, Forest Service. (2000) Cave Creek Ranger District
OHV Map: Tonto National Forest, Arizona. Washington, DC: Government Printing Office.
U.S. Department of the Interior, Bureau of Land Management. (1987) Recreation
Area Management Plan and Environmental Assessment for the Imperial Sand
Dunes. Washington, DC: Government Printing Office.
U.S. Department of the Interior, Bureau of Land Management. (1989) Off-Road
Vehicle Recreation: Buffalo Resource Area. Washington, DC: Government Printing Office.
Vogler, J. B. and Butler, D. R. (1996) Pedestrian- and bicycle-induced path erosion
on a university campus. Physical Geography, Vol. 17, 485–494.
Vollmer, A. T., Maza, B. G., Medica, P. A., Turner, F. B., and Bamberg, S. A. (1976)
The impact of off-road vehicles on a desert ecosystem. Environmental Management, Vol. 1, 115–129.
Wallin, T. R. and Harden, C. P. (1996) Estimating trail-related soil erosion in the
humid tropics: Jatun Sacha, Ecuador, and La Selva, Costa Rica. Ambio, Vol. 25,
517–522.
Weaver, T. and Dale, D. (1978) Trampling effects of hikers, motorcycles and horses
in meadows and forests. Journal of Applied Ecology, Vol. 15, 451–457.
Webb, R. H. and Wilshire, H. G., eds. (1983) Environmental Effects of Off-Road
Vehicles: Impacts and Management in Arid Regions. New York, NY: SpringerVerlag.
Whinam, J., Cannell, E. J., Kirkpatrick, J. B., and Comfort, M. (1994) Studies on the
potential impact of recreational horseriding on some alpine environments of the
central plateau, Tasmania. Journal of Environmental Management, Vol. 40, 103–
117.
Williams, T. (2000) Motorizing public land. Audubon, Vol. 102, 42–52.
Wilshire, H. G., Nakata, J. K., Shipley, S., and Prestegaard, K. (1978) Impacts of
vehicles on natural terrain at seven sites in the San Francisco Bay area. Environmental Geology, Vol. 2, 295–319.
Wilson, J. P. and Seney, J. P. (1994) Erosional impact of hikers, horses, motorcycles,
and off-road bicycles on mountain trails in Montana. Mountain Research and
Development, Vol. 14, 77–88.
Ziegler, A. D., Sutherland, R. A., and Giambelluca, T. W. (2000) Runoff generation
and sediment production on unpaved roads, footpaths and agricultural land surfaces in northern Thailand. Earth Surface Processes and Landforms, Vol. 25,
519–534.
Ziegler, A. D., Sutherland, R. A., Giambelluca, T. W. (2001) Acceleration of
Hortonian flow and erosion by footpaths in an upland agricultural watershed in
northern Thailand. Geomorphology, Vol. 41, 249–262.