Forgot your password?
Document Actions
McDonnell 1981
Btological Conservation 21 (1981)289 301
TRAMPLING EFFECTS ON COASTAL DUNE
VEGETATION IN THE PARKER RIVER NATIONAL
WILDLIFE REFUGE, MASSACHUSETTS, USA
M. J. MCDONNELL*
Department of Botany and Plant Pathology, University of New Hampshire,
Durham, New Hampshire 03824, USA
ABSTRACT
Long-term human trampling on coastal dune vegetation was analysed using eight
100 × 100m study areas ranging in trampling intensity JJ'om little or no past
trampling pressure to severe trampling pressure. All trampling levels significantly
reduced species diversity. Total vegetation cover was significantly reduced only in
areas with severe and heavy trampling pressure. Where long-term moderate
trampling did not affect total cover, it appears to have Javoured Ammophila
breviligulata over other more sensitive species. The data suggest that long-term
trampling pressure has caused a shift in the horizontal pattern oJ vegetation, i.e.
expanding the width of thejbredune plant community, while restricting the interdune
community to a narrow strip at the base of" the backdune ridge.
INTRODUCTION
Vegetation adapted to the dune environment is continually exposed to strong winds,
moving sand and salt spray. The plant cover is often thin or sparse and forms a
fragile network holding the sand in place. The dune system, though well adapted to
absorbing the impact of coastal storms, is fragile and can be severely damaged by
livestock, pedestrian and vehicular traffic (Clark, 1974; Olsen & Grant, 1975). If a
multiple-use approach is employed for coastal dunes, management guidelines must
be developed to ensure preservation of these fragile systems.
Much descriptive work has been done regarding dune vegetation and factors
affecting its distribution (Oosting & Billings, 1942; Martin, 1959; Willis et al.,
1959a,b; van der Valk, 1974, 1975; Lamoureux & Grandtner, 1977). Within the last
* Present address: Department of Botany, Rutgers University, Piscataway, New Jersey 08854, USA.
289
Biol. Conserr. 0006-3207/81/0021-0289/$02'75 ~: Applied Science Publishers Ltd, England, 1981
Printed in Great Britain
290 M.J. McDONNELL
ten years research, particularly in Europe, has been conducted to assess the human
impacts on dune systems (Burden & Randerson, 1972; Oldfield et al., 1972; Trew,
1973; Liddle, 1975a,b; Liddle & Greig-Smith, 1975a,b; Blom, 1977; Boorman &
Fuller, 1977; Richards & Stead, 1978). Along the New England coast (USA)
most research of this nature specifically concerns the effects of vehicle traffic on the
dunes (Niedoroda & Limeburner, 1975; Broadhead & Godfrey, 1977). Nickerson
(1976) studied the impact of human trampling on the dunes in Cape Cod National
Seashore and found that vulnerability of the predominant foredune species,
Ammophila breviligulata, to foot traffic varied with time of growing season, and that
shoe traffic was 10 to 20 times more destructive than bare-foot traffic.
Studies by Liddle & Greig-Smith (1975b) have shown that wear by vehicles and
human trampling on sand dunes reduces species diversity as well as population size
and distribution. They conclude that the consequences of this type of impact are as
great as the major natural environmental factors.
The majority of research on the effect of human trampling on coastal dunes has
been directed toward the response of the vegetation (i.e. specific species)
immediately exposed to the impacts. This type of impact research has also been the
major focus in other ecosystems (LaPage, 1967, 1968; Goldsmith et al., 1970;
Willard & Marr, 1970, 1971; Dale & Weaver, 1974; Foin et al., 1977). In
contrast, a second approach which examines the distribution and species
composition of plant communities which exist in areas which have had long-term
trampling pressures and adjacent areas with little or no trampling pressure provides
an assay of the effect of long-term trampling on vegetation patterns.
The objective of this study is to determine the effect of long-term trampling
pressure on the composition and horizontal pattern of coastal dune vegetation in the
Parker River National Wildlife Refuge.
THE STUDY AREA
The site selected was the dune system on Plum Island, Essex County, Massachusetts
(Fig. 1). The island and its flora were described by McDonnell (1979a,b). Two-thirds
of this 12.8km-long by approximately 0.6km-wide barrier island has remained
relatively natural. About 1,230 ha of the southern portion of the island are now
included in the Parker River National Wildlife Refuge and a Commonwealth of
Massachusetts State Park, which owns the extreme southern tip. This is one of the
largest semi-natural barrier beach dune systems in New England north of Cape Cod.
Plum Island formed after the retreat of the Wisconsin Ice Sheet some 6,300 years
ago (Mclntire & Morgan, 1964). It is composed primarily of unconsolidated beach
and dune sands and marsh deposits (Jones & Cameron, 1976). The southern two-
thirds of the island exhibits classical features of a barrier beach dune system. Four
TRAMPLING EFFECTS ON COASTAL DUNE VEGETATION 29
~er
~ sland
N i ~ '--~ RJ.
co Atlantic Ocean
(¢
Plum Island ,o
Sound
0
Z
"7
State P a r k -~-.--d metres
0 300 900
Fig. 1. A m a p of Plum Island showing the location of the study areas in the Parker River Nationa
Wildlife Refuge, Insert is a m a p of New England showing the location of Plum Island.
292 M.J. McDONNELL
major physiographic zones can be identified running east to west across the island
(Fig. 2): (1)beach; (2)foredune; (3)interdune; and (4)backdune.
The adjacent ocean is an important moderating factor in the climate. The average
annual temperature for this region is 10 °C with a January average of - 1.4 °C and a
July average of 19.9 °C. Average annual precipitation is 193.1 cm with the winter
season being the wettest due to frequent storms. Winds are primarily westerly or
offshore. The storm winds on the other hand are most prevalent from the east, with
those most important to beach formation coming from the northeasterly direction.
BEACH FOREDUNE INTERDUNE = BACKDUNE
L Tr ~ '
Ac
.c ~qt~
~'I
Ps
~'-X RC
Ab LI . Ht ~ , ~
A.
Co
Pm Ac Ps
Ss Mp /
B.
Fig. 2. A diagrammatic sketch of the typical physiographic features and vegetation zones encountered
on two east to west transects across Plum Island: A, area of the refuge with little or no trampling pressure;
B, area exposed to heavy trampling pressure. (Key: Ab, Ammophila breviligulata: Ac, Amelanchier
canadensis: Co, Cehis occidentalis: Cp, Carex pennsylvanica: Ht, Hudsonia tomentosa; Lj, Lathyrus
japonicus; Mp, Myrica pennsylvanica; Pm, Prunus maritima; Ps, Prunus serotina: Rc, Rosa carolina;
Ss, Solidago sempervirens; and Tr. Toxicodendron radicans).
Because of the proximity of Plum Island to a large metropolitan area (Boston),
the refuge management has carefully regulated the number of vehicles entering the
refuge area, particularly during heavy-use periods. Access to the dunes has been
limited to ten parking areas spaced at intervals along the 9.7kin road to the
southern end of the island. Thus trampling pressure has been concentrated in
specific well-defined areas of the dune system. In addition, to limit trampling on the
dune vegetation adjacent to these parking areas, boardwalks to the beach were
established over the past six years. The refuge management has also set aside a large
Dune Natural Area for research which receives very little trampling pressure. Since
the late 1960s the refuge management has not allowed vehicles on the dunes; thus
T R A M P L I N G EFFECTS O N C O A S T A L D U N E V E G E T A T I O N 293
any trampling is caused by either humans or wildlife (e.g. dogs, rabbits, etc.). Plant
nomenclature follows Fernald (1950). Voucher specimens of plants were collected
and deposited in the Hogdon Herbarium, University of Hampshire (NHA).
METHODS
A preliminary examination of the impact of human trampling on the dune
vegetation in the Parker River National Wildlife Refuge revealed that the foredune
and interdune areas were the most severely impacted. These areas are primarily
covered with perennial herbs, e.g. Ammophila breviligulata and Lathyrusjaponicus.
while the backdune vegetation is dominated by woody trees and shrubs. A t'e~.
distinct paths extend through the tree and shrub colonies surrounding each parking
area but it is not until the path reaches the open interdune and foredune areas that
large-scale destruction and even elimination of vegetation can be observed.
To obtain quantitative information on the horizontal structure of the vegetation
in heavily trampled and relatively untrampled foredune and interdune areas, eight
study areas were established. Five study areas were adjacent to parking lots.
alternating among the ten which provide access across the dunes. Study areas were
established adjacent to parking lots 1 (P 1), 3 (P3), 5 (P5), 8 (P8) and 10(P10) (Fig. l).
The study areas were then grouped according to trampling intensity. Accurate
estimates of the actual number of visitors or cars present at a parking lot over a
season are not available. A request by the author to distribute questionnaires to
determine parking lot use patterns was denied due to refuge management policy.
Visual observations of each parking lot over an extended time were beyond the
resources of this project. However, it is felt that, because the emphasis of the study is
on long-term trampling impacts on the area, current parking lot use patterns
obtained from first-hand observations and information from refuge personnel.
along with knowledge of the history and size of each parking lot+ provide an
adequate estimate of trampling intensity. Thus, estimates of trampling intensity for
each parking lot study area were based on (l)current parking lot use trends~
(2) capacity of parking lot to holds cars; (3) number of years 9pen; (4) distance trom
refuge entrance; and (5) length of time boardwalks were established.
Over the past ten years there has been a dramatic increase in the number of people
using the refuge. During a single year, from October 1976 to September 1977. ~ total
of 469,482 people visited the refuge (Parker River National Wildlife Refuge
records). The greatest number occurs during the summer months with an average of
over 50,000 people per month. With only one entrance to the refuge, at the northern
end of Plum Island, there is a distinct gradient of recreational use down the island.
The closer the parking lot is to the entrance the greater use it gets. Thus, parking lots
]-6 get many more visitors than parkings lots 7 11. In general most people do not
drive 9.6 km down a dirt road to the parking lots at the southern end of the island.
From May through August when the weather is pleasant, all of the parking lots
294 M.J. McDONNELL
are filled close to capacity. In the last few years they also fill up rapidly during the
autumn months ( P R N W R records). With such heavy use the capacity of the parking
lots provides a good estimate of the intensity of trampling on the adjacent study
area. The smaller the parking lot, the less intense the trampling and vice versa.
Thus, an area of dune adjacent to a small parking lot at the southern end of the
refuge is less intensely trampled than an area adjacent to a large parking lot at the
northern end of the refuge.
Using this information, coupled with the number of years a parking lot has been
open and the length of time it has had a boardwalk, the parking lot study areas were
grouped into three categories according to level of probable trampling pressures:
severe P1; heavy P3, P5; and moderate P8, PI0 (Table 1).
TABLE 1
PARKING LOT STUDY AREAS GROUPED INTO THREE CATEGORIES OF PROBABLE TRAMPLING PRESSURE
ACCORDING TO AGE, SIZE, DISTANCE FROM REFUGE ENTRANCE AND NUMBER OF YEARS BOARDWALKS WERE
PRESENT
Approximate Distance Total years Estimated
Study Parking Capacity no. of years from refuge boardwalk trampling
area lot (no. of cars) open entrance present pressure
P1 1 145 30 0.09 k m 2 Severe
P3 3 31 28 0.70 k m 6 Heavy
P5 5 22 24 2.08 k m 6 Heavy
P8 8 9 13 6.40 k m 3 Moderate
P 10 10 26 13 8-80 k m 3 Moderate
In addition, three study areas (N1, N2 and N3) were established in relatively
undisturbed areas as far from potential trampling pressure as possible (Fig. 1). Two
study areas were established in the Dune Natural Area: N1 approximately 1200 m
south of parking lot 5; N2 approximately 1600 m north of parking lot 6. Study area
N3 was established in a relatively undisturbed and inaccessible area of the dunes
between parking lots 8 and 9.
Each study area was 100 x 100m with the edges oriented north-south and
east-west. The eastern edge of a study area was the foredune ridge. The study areas
adjacent to parking lots were set up so that the midpoint of the eastern boundary was
the intersection of the eastern end of the boardwalk and the foredune ridge. Those
study areas not adjacent to parking lots were established so that the midpoint of the
eastern boundary was a point originally established on a map and then located in the
field. In each study area four 10 x 100 m transects running north south were centred
on 10, 35, 65 and 95 m respectively from the eastern edge of the foredune ridge. Each
transect was further divided into 40 5 x 5 m quadrats, situated as pairs on either
side of a centre line. Due to the fragile nature of the dune system and refuge
management policy, destructive biomass sampling was not attempted. Instead,
percent cover and frequency values were obtained to determine species composition
and abundance of the vegetation in the study areas.
T R A M P L I N G EFFECTS O N C O A S T A L D U N E V E G E T A T I O N 295
During August 1978 20 randomly chosen 5 x 5 m quadrats in each transect were
sampled for percent cover of species present, total number of species present, and
total area with vegetation cover. Quadrats were chosen at random because the
position of the boardwalks shifts from year to year within about a 50-m area. Thus,
no attempt was made to quantify the gradient of trampling away from the
boardwalks since the exact intensity of trampling at any one point has varied over
the years depending on the position of the boardwalks. The data were statistically
analysed using analysis of variance techniques. Because the quadrats sampled were
randomly chosen they could not be pooled to form the experimental error term for
they may not have the same variance as postulated in the mathematical model. To
provide a more rigorous test the percent cover values were averaged over transects,
and an error term was constructed using selected single degree freedom sums of
squares for transect and location contrasts, set up prior to data analysis. The exact
procedure used follows Cochran & Cox (1957) and is described in McDonnell
(1979a). After completion of the analysis of variance, treatment means were
compared using Least Significant Difference (LSD) at alpha 0-05.
RESULTS A N D D I S C U S S I O N
Twenty-five species were present in the eight study areas. Ammophila breviligulata,
Lathyrus japonicus, Artemisia stellariana and Hudsonia tomentosa together
account for over 75 ~o of the total area covered with vegetation in the study areas
(Table 2). The percent cover of the less dominant species is presented in Table 3.
TABLE 2
THE PERCENT COVER OF THE DOMINANT SPECIES AND TOTAL AREA COVERED WITH VEGETATION AND THE
A V E R A G E NUMBER OF SPECIES PRESENT IN EACH S T U D Y AREA
L S D , L e a s t Significant Difference. In the a n a l y s i s o f v a r i a n c e t h e r e were f o u r t r a n s e c t s a n d eight
l o c a t i o n s ; in n o case w a s the t r a n s e c t X l o c a t i o n i n t e r a c t i o n significant, F-test significant at 0.05 level.
Species Stud)' areas
PI P3 P5 P8 PIO NI N2 N3 ANOVA +
Trampling Average LSD Transect Location
pressure: severe heavy moderate undisturbed
Ammophila
breviligulata 6-7 18.0 20.6 30'6 22.5 18-0 18-8 25.0 20"0 8.5 NS °
Lathyrus
japonicus 2.8 9-9 2.4 12.4 3.2 9.0 4-9 9.0 6.7 6.1 " NS
A rtemisia
stellariana 0" 1 2.0 1-4 2.6 1 "2 7-6 1 "2 0"0 2.0 9.7 NS NS
Hudsonia
tomentosa 0.0 0.5 0.3 0-2 5-3 3-0 8.6 9.0 3.3 6-1 ° "
Total
v e g e t a t i o n cover 9.6 36.0 25.8 46.0 39.0 40-0 43.9 53.4 38.4 14-2 NS "
Average
n o . o f species 0.6 2.0 2.5 2.7 2-8 3.2 3.4 4.0 2.7 0.46 "
N S = n o t significant. " = significant.
296 M . J . McDONNELL
TABLE 3
THE PERCENT COVEROF THE LESSDOMINANTSPECIESWHICH OCCUR IN THE EIGHTSTUDYAREASLOCATEDIN
THE PARKER RIVER NATIONALWILDLIFEREFUGE
Trampling pressure: PI, severe; P3 + P5, heavy; P8 + P10, moderate; NI, N2 and N3, undisturbed.
Stud), areas
Species PI P3 P5 P8 P10 N1 N2 N3
Solidago sempervirens 0"00 0.31 0.41 0.80 1.03 2.34 1.51 0.65
Euphorbiapolygonifolia 0"02 0.03 0.43 0.03 0.18 0.10 0.28 0.08
Cyperus sp. 0.00 0"00 0.00 0.00 0.07 0.13 0.11 1.14
Polygonella articulata 0"00 0"02 0.00 0.00 0.06 0.10 0.14 0.76
Cakile edentula 0.00 0.04 0.10 0.38 0.03 0.00 0-03 0-02
Rosa sp. 0-00 0.87 0-00 0.00 0.13 0.00 3-00 0.00
Prunus serotina 0"00 0.00 0.00 0"00 0.06 0"00 0.00 0.00
Polygonum scandens 0.00 0.00 0-00 0-00 0-01 0-00 0-00 0.00
Spiraea tomentosa 0.00 0.00 0.00 0"00 0"00 0.00 0.43 0.00
Viburnum recognitum 0'00 0'00 0"00 0"00 0"00 0"00 0-43 0'00
Fraxinus pennsylvanica 0-00 0.00 0.00 0.00 0.00 0-00 0.25 0"00
Xanthium eehinatum 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00
Toxicodendron radicans 0.00 4.18 0.00 0.00 1.62 0.00 2.81 0-00
Myrica pennsylvanica 0-00 0-80 0-00 0-00 4.00 0-00 1-43 0-00
Raphanus raphanistrum 0-00 0'00 0-00 0.18 0"00 0"00 0-00 0"00
Prunus maritima 0.00 0.00 0.00 0.00 0.00 0.26 0.04 0.00
Lechea maritima 0"00 0"00 0"00 0"00 O.13 0"00 0"00 O"18
Spartina patens 0.01 0"00 0-00 0-00 0.00 0.00 0.00 0.00
Distichlis spicata 0.01 0'00 0"00 0"00 0'00 0'00 0.00 10.00
Cladonia sp. 0.00 0.00 0.00 0.00 0.44 0"00 0"00 4.50
Geaster sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.03
An analysis of variance (Table 4) of the percent cover of the total area covered
with vegetation revealed that differences among transects (variation within study
areas) were not significant, whereas differences between locations (different
trampling pressures) were. The parking lot study areas as a group have significantly
less vegetation cover than the undisturbed study areas. When the means for each
study area are compared (Table 2) only parking lot study areas PI and P5 (with
severe and heavy trampling pressure respectively) exhibited significantly less
vegetation cover than the undisturbed study areas (N1, N2 and N3). Moderately
trampled parking lot study areas P8 and P10 had significantly more vegetation than
study areas P1 and P5 but were not significantly different from the undisturbed
study areas or study area P3 (heavy trampling pressure). The relatively high cover
value for study area P3 results partially from the fact the boardwalk was moved
50 m north of its previous location just before the study began, thus excluding some
of the more heavily impacted area.
The most important species stabilising the foredune and interdune zones is
A. breviligulata. The analysis of variance (Table 4) indicates that, like total
vegetation cover, the only significant factor was location (i.e. different trampling
pressures). But unlike the total vegetation cover analysis, the undisturbed compared
with disturbed study areas contrast was not significant. Only severely trampled
study area P1 had significantly less percent cover of A. breviligulata than the
undisturbed study areas (Table 2). Study areas with heavy and moderate trampling
TRAMPLING EFFECTS ON COASTAL DUNE VEGETATION 29"]
TABLE 4
A N A L Y SI S OF V A R I A N C E TABLE F O R THE TOTAL AREA COVERED WITH VEGETATION AND Ammophila
breviligulata
Four transects were sampled at each of the eight study areas. N, undisturbed study areas, N l, N2, N 3; P.
parking lot study areas, PI has severe trampling pressure, P3 and P5 have heavy trampling pressure, and
Pg and PI0 have moderate trampling pressure.
Total vegetation cover Ammophila breviligulata
Sums Mean Sums Mean
Sources of variation DF of squares square oj squares square
Overall mean 1 872796 258898
Transect (T) 3 6484 2161 798 266
Location (L) 7 103880 14840 ~ 26322 3760"
LI (N vs. P) 1 32374 32374 ~ 128 128"
L2 (N1 vs. N2) 1 354 354 12 12
L3 (NI + N2 vs. N3) l 6468 6468 2170 2170
L4 (P8 vs. PI0) 1 1904 1904 2621 2621
L5 (P3 vs. P5) 1 4270 4270 228 228
L6 (P8 + P10 vs. P3 + P5) 1 10981 10981" 4039 4039
L7 (P1 vs. P3 + P5 + P8 + PI0) 1 47517 47517 ~ 17125 17125"
L x T 6 7914 1319 7896 1316
Experimental error 15 26378 1758 14184 945
" Significant: F-test, alpha = 0.05.
pressure (P3, P5 and P8, P10, respectively) did not have significantly less cover of
A. breviligulata (Table 2). In fact, study area P8, with moderate trampling pressure,
had significantly higher percent cover of A. breviligulata than undisturbed study
areas N 1 and N2. In the trampled parking lot study areas A. breriligulata made up
50',~;~ 80 ~'0 of the total vegetation cover while in the undisturbed study areas it
o, o/
contributed only 42/o 47/,, to the total cover. It appears that long-term moderate
trampling pressure does not affect total cover, but favours A. hreviligulata over
other more sensitive species.
An analysis of variance of the percent cover of the other dominant species.
L.japonicus, A. stellariana and H. tomentosa, revealed that only the last could he
used to estimate the effect of trampling. Location (different trampling pressures) had
no significant effect on the total cover of L.japonicus or A.,stellariana (Table 2).
Both of these species occur sporadically throughout the dune system.
An AN OVA table for the percent cover of H. tomentosa (Table 5) indicates that,
unlike the preceding species, both transect and location factors were significant. The
percent cover of H. tomentosa was significantly greater in transects C and D, which
are farthest from the foredune, than A and B, which are on the foredune. This species
commonly forms large mats behind the foredune in the stable protected areas of the
interdune (Fig. 2). When the percent cover means for each study area are compared
(Table 2), the undisturbed study areas exhibit significantly more H. tomenlosa than
any of the trampled study areas except PI 0. which has moderate trampling pressure.
No H. tomentosa was present in the severely trampled study area P1, while in heavily
and moderately trampled study areas P3, P5 and P8 respectively, it never occurred
in amounts greater than 1 '~, cover.
298 M.J. McDONNELL
TABLE 5
ANALYSIS OF VARIANCE TABLE FOR THE PERCENT COVER OF nudsonia tomentosa AND FOR THE AVERAGE
NUMBER OF SPECIES PRESENT IN THE EIGHT STUDY AREAS
Hudsonia tomentosa Average number o.] species
Sums Mean Sums Mean
Source o]variation DF of squares square DF of squares square
Overall m e a n 1 7388 1 4665
Transect(T) 3 5788 192 a 3 129 434
L o c a t i o n (L) 9 8249 1178 a 7 576 82 a
T × L 9 4955 550 6 11 2
Experimental error 12 5802 483 15 98 6
Significant; F-test, a l p h a = 0.05.
The increased importance of A. breviligulata and the virtual elimination of
H. tomentosa in areas of the dune exposed to trampling pressure suggest that
trampling has caused the typical foredune community dominated by
A. breviligulata to extend inland, reducing the width of the interdune community.
Thus, the interdune community has become restricted to a narrow strip near the base
of the backdune ridge (Fig. 2).
In general, plant species diversity in the foredune and interdune regions of the
dune system is low. An analysis of variance on the average number of species present
in the study areas revealed that both transect and location factors were significant
(Table 5). Transects farthest from the foredune (C&D) exhibited the greatest
species diversity. A comparison of the average number of species in each study area
(Table 2) revealed that a significantly greater number of species occurred in the
undisturbed study areas as opposed to the study areas exposed to trampling
pressure (average 3.5 spp. area in the undisturbed sites compared with 2.3 spp. area
in areas with heavy trampling pressure). Thus, long-term trampling pressure, in
addition to changing the horizontal pattern of the coastal dune vegetation, has also
reduced species diversity.
CONCLUSION
The effects of long-term human trampling on the coastal dune vegetation in the
Parker River National Wildlife Refuge varied depending on the intensity of
trampling pressure. Severe trampling pressure caused a drastic reduction in both
vegetation cover and species diversity (Fig. 3). Somewhat less conclusive, heavy
trampling pressure appeared to cause a reduction in cover and diversity. Moderate
trampling pressure significantly reduced species diversity, but not total cover. Long-
term moderate trampling appears to have favoured A. breviligulata over other more
sensitive species. In contrast, all levels of trampling pressure significantly reduced
the amount of H. tomentosa present. Even though moderate long-term trampling
pressure has not caused significant and obvious changes in vegetation cover, similar
T R A M P L I N G EFFECTS ON C O A S T A L D U N E V E G E T A T I O N 299
80' "4
70 Average <~
Number of S p e c i e s \
\
w
\
\
60 \ 3
\
a
\
\
50 \.
w
\
,ota, \
"4'
Vegetation Cover ~~ /~
\ 2 w
40 \
\ \
\
\
O
30
\
m
J
AmmoDhila [ ~
Z
20 brevilioulata
\ • <
10
Hudsonia
tomentosa
I I
UNDISTURBED MODERATE HEAVY SEVERE
N1, N2, N3 P8, P10 P3, P5 Pt
TRAMPLING PRESSURE
Fig. 3. A summary of the effects of different trampling pressures on the coastal dune vegetation in the
Parker River National Wildlife Refuge. N and P refer to numbered study areas. Solid symbols indicate a
significant difference from undisturbed study areas at alpha = 0.05.
to severe and heavy trampling, it has caused a loss o f species, a shift in relative
a b u n d a n c e o f species, and a shift in the horizontal pattern o f the vegetation. The
increased importance o f A. breviligulata, and the virtual elimination o f interdune
species such as H. tomentosa in all areas exposed to trampling pressure, indicates
that long-term h u m a n trampling has resulted in the expansion o f the foredune
c o m m u n i t y d o m i n a t e d by A. breviligulata and the restriction o f the interdune
c o m m u n i t y to a narrow strip at the base o f the backdune ridge (Fig. 2).
300 M . J . McDONNELL
Smaller p a r k i n g lots a n d the e s t a b l i s h m e n t of boardwalks has successfully
reduced the impact of t r a m p l i n g on the vegetation in the foredune a n d i n t e r d u n e
regions in the Parker River N a t i o n a l Wildlife Refuge. Results from this study
suggest that destruction of d u n e vegetation by t r a m p l i n g can be minimised if d u n e
access is restricted to specific areas with p a r k i n g lots designed to hold only 10 to 20
cars a n d have b o a r d w a l k s that extend over the d u n e s to the beach.
ACKNOWLEDGEMENTS
I would like to express my gratitude to Dr G a r r e t t E. Crow and D r Russell K i n e r s o n
for their help t h r o u g h o u t this study. I wish to t h a n k George Gavutis, m a n a g e r of the
Parker River N a t i o n a l Wildlife Refuge: w i t h o u t his c o o p e r a t i o n this study could n o t
have been completed. The following persons assisted with the field work: Lisa
B a n d a z i a n , L y n n Bohs, Allison Buck, Bill C o t r e a u a n d Dennis M c K e a r i n . Dr Jerry
W a r r e n a n d the staff of the office of Academic C o m p u t i n g at U N H provided
valuable assistance in the statistical analysis of the data. I would also like to t h a n k
Dr Richard T. T. F o r m a n and an a n o n y m o u s reviewer for their constructive criticism
of the m a n u s c r i p t . This study was supported in part by two Central University
Research G r a n t s ( C U R F $86 a n d S106) from the University of New Hampshire.
REFERENCES
BLOM, C. W. P. M. (1977). Effect of trampling on soil compaction on the occurrence of some Plantago
species in coastal sand dunes, II. Trampling and seedlingestablishment. Oecol. Plant., 12, 363 81.
BOORMAN, L. A. & FULLER, M. (1977). Studies on the impact of paths on the dune vegetation at
Winterton, Norfolk, England. Biol. Conserr., 12, 203 16.
BROADHEAD,J. M. & GODFREY,P. J. (1977). Off road vehicle impact in Cape Cod National Seashore:
Disruption and recovery of dune vegetation. Int. J. Bioclim. Biomet., 21,299 306.
BURDEN.R. F. & RANDERSON,P. F. (1972). Quantitative studies of the effects of human trampling on
vegetation as an aid to the management of semi-natural areas. J. appl. Ecol., 9, 439 58.
CLARK, J. (1974). Coastal ecosystems. Washington, DC, Conservation Foundation.
COCHRAN,W. & Cox, G. (1957). Experimental design. New York, John Wiley.
DALE, D. & WEAVER,T. (1974). Trampling effects on vegetation of the trail corridors of north Rocky
Mountain forests. J. appl. Ecol., 11, 767-72.
FERNALD,M. L. (1950). Gray's manual oJbotany, 8th edn, corrected edn, 1970. New York, Van Nostrand
Reinhold.
FOIN,T. C., GARTON,E. O., BOWEN,C. W., EVERINGHAM, M., SCHULTZ,R. O. & HOLTON,G. (1977).
J.
Quantitative studies of visitor impacts on environments of Yosemite National Park, California, and
their implications for park management policy, J. environ, mgmt, 5, 1 22.
GOLDSMITH,F. B., MUNTON,R. J. C. & WARREN,A. (1970). The impact of recreation on the ecologyand
amenity of semi-natural areas: methods of investigation used in the Isles of Scilly. Bin. J. Linn. Soc.,
2, 287 306.
JONES,J. R. & CAMERON,B. (1976). Sedimentaryand geomorphicorigin and developmentof Plum Isla.nd,
Massachusetts: An example of a barrier island system. In Geology oJSouthern New England. A
guidebook jor field trips to the Boston area and vicinity, ed. by B. Cameron, 188 204. Boston, Science
Press.
TRAMPLING EFFECTS ON COASTAL DUNE VEGETATION 301
LAMOUREUX,G. & GRANDTNER,M. M. (1977). Contribution/~ l'etude ecologique des dunes mobiles, 1
Les 61~ments phytosociologique. Can. J. Bot., 55, 158 71.
LAPAGE, W. P. (1967). Some observations on campground trampling and ground cover response. U.S
Forest Service Res. Paper N. E. 68.
LAPAGE, W. P. (1968). Recreation and the forest site. J. For., 66, 319 21.
LIDDLE, M. J. (1975a). A selective review of the ecological effects of human trampling on natural
ecosystems. Biol. Conserv,, 7, 17-36.
L1DDLE, M. J. (1975b). A theoretical relationship between the primary productivity of vegetation and its
ability to tolerate trampling. Biol. Conserv., 8, 251 5.
LIDDLE, M. J. & GREIG-SMITH, P. (1975a). A survey of tracks and paths in a sand dune ecosystem. 1. Soils.
J. appl. Ecol., 12, 893 908.
LIDDLE, M. J. & GREIG-SMITH, P. (1975b). A survey of tracks and paths in a sand dune ecosystem. 11
Vegetation. J. appl. Ecol., 12, 909 30.
MARTIN, W. E. (1959). The vegetation of Island Beach State Park, New Jersey. Ecol. Monogr., 29.1 46
McDONNELL, M. J. (1979a). The vascular flora oJ Plum Island. Essex Count.l', Massachus'ett.s. with an
analysis of the impact oJhuman trampling on coastal dune regetation. MS thesis, University of New
Hampshire, Durham, NH.
MCDONNELL, M. J. (1979b), The flora of Plum Island, Essex County. Massachusetts. Bull. New Hamp~
agric. Exp. Stn, 513.
MCINTYRE, W. & MORGAN, J. (1964). Recent geomorphic history oj Plum Island, Massachu.sett.s. and
adjacent coasts. Baton Rouge, Louisiana State University Press.
NICKERSON, N. (1976). Effects of pedestrian traffic on sand dune vegetation. In Abstract,s ./rom
ConJerence on Scientific Research in the National Parks. 1st. New Orleans, Louisiana. American
Institute of Biological Sciences and National Parks Service. United States Department of the
Interior.
N1EDORODA,A. & LIMEBURNER,R. (1975). A preliminary report on the geomorphic effects of recreational
vehicles on coastal systems. National Park Service Cooperative Park Studies Unit Rept., No 7
Amherst, Ma.. University of Massachusetts.
OLDFIELD, F., CARTER,R., KITCHER,K. & WILCOCK. F. (1972). Report oj investigations into coastal
erosion and accretion along the coastlines oj County Antrim and County Londonderry. ('oasta/
Research Project Report 196~1972. School of Biological and Environmental Studies. New
University of Ulster.
OLSEN, S. B. & GRANT, M. J. (1975). Rhode Island's Barrier Beaches: Volume 1. A report on a
management problem and an evaluation of options. A Coastal Management Publication
University of Rhode Island. Marine Technical Report No.4. Kingston. R.I.
OOSTING, H. J. & BILLINGS, W. D. (1942). Factors effecting vegetational zonation on coastal dunes.
Ecology, 23, 131-42.
RICHARDS, 1. & STEAD, F. (1978). Dune conservation in Anglesey. Tn Ctrv Plann., 46, 173 6.
TREW, M. J. (1973). The effects and management of trampling on coastal sand dunes, d. Environ. Plann. &
Pollut. Control. 1, 38-49.
VAN DER VALK,m. G. (1974). Mineral cycling in coastal foredune plant communities in Cape Hatteras
National Seashore. Ecology, 55, 1349 58.
VAN DER VALK, A. G. (1975). The floristic composition and structure of foredune plant communities of
Cape Hatteras National Seashore. Chesapeake Sci., 16, 127 33.
WILLARD, B. & MARR, J. (1970). Effects of human activities on alpine tundra ecosystems in Rock 5
Mountain National Park, Colorado. Biol. Conserv., 2, 257-65.
WILLARD, B & MARR, J. (1971). Recovery of alpine tundra under protection after damage by human
activity in the Rocky Mountains of Colorado. Biol. Conserv., 3, 131 90.
WILLIS, A. J., FOLKES, B. F., HOPE-SIMPSON,J. F. & YEMM, E. W. (1959a). Braunton Burrows: The dune
system and its vegetation, Part I. J. Ecol.. 47, 1 22.
WILLIS, A. J.. FOLKES,B. F., HOPE-SIMPSON,J. F. & YEMM, E. W. (1959b). Braunton Burrows: The dune
system and its vegetation, Part II. J. Ecol.. 47, 249 87.
TRAMPLING EFFECTS ON COASTAL DUNE
VEGETATION IN THE PARKER RIVER NATIONAL
WILDLIFE REFUGE, MASSACHUSETTS, USA
M. J. MCDONNELL*
Department of Botany and Plant Pathology, University of New Hampshire,
Durham, New Hampshire 03824, USA
ABSTRACT
Long-term human trampling on coastal dune vegetation was analysed using eight
100 × 100m study areas ranging in trampling intensity JJ'om little or no past
trampling pressure to severe trampling pressure. All trampling levels significantly
reduced species diversity. Total vegetation cover was significantly reduced only in
areas with severe and heavy trampling pressure. Where long-term moderate
trampling did not affect total cover, it appears to have Javoured Ammophila
breviligulata over other more sensitive species. The data suggest that long-term
trampling pressure has caused a shift in the horizontal pattern oJ vegetation, i.e.
expanding the width of thejbredune plant community, while restricting the interdune
community to a narrow strip at the base of" the backdune ridge.
INTRODUCTION
Vegetation adapted to the dune environment is continually exposed to strong winds,
moving sand and salt spray. The plant cover is often thin or sparse and forms a
fragile network holding the sand in place. The dune system, though well adapted to
absorbing the impact of coastal storms, is fragile and can be severely damaged by
livestock, pedestrian and vehicular traffic (Clark, 1974; Olsen & Grant, 1975). If a
multiple-use approach is employed for coastal dunes, management guidelines must
be developed to ensure preservation of these fragile systems.
Much descriptive work has been done regarding dune vegetation and factors
affecting its distribution (Oosting & Billings, 1942; Martin, 1959; Willis et al.,
1959a,b; van der Valk, 1974, 1975; Lamoureux & Grandtner, 1977). Within the last
* Present address: Department of Botany, Rutgers University, Piscataway, New Jersey 08854, USA.
289
Biol. Conserr. 0006-3207/81/0021-0289/$02'75 ~: Applied Science Publishers Ltd, England, 1981
Printed in Great Britain
290 M.J. McDONNELL
ten years research, particularly in Europe, has been conducted to assess the human
impacts on dune systems (Burden & Randerson, 1972; Oldfield et al., 1972; Trew,
1973; Liddle, 1975a,b; Liddle & Greig-Smith, 1975a,b; Blom, 1977; Boorman &
Fuller, 1977; Richards & Stead, 1978). Along the New England coast (USA)
most research of this nature specifically concerns the effects of vehicle traffic on the
dunes (Niedoroda & Limeburner, 1975; Broadhead & Godfrey, 1977). Nickerson
(1976) studied the impact of human trampling on the dunes in Cape Cod National
Seashore and found that vulnerability of the predominant foredune species,
Ammophila breviligulata, to foot traffic varied with time of growing season, and that
shoe traffic was 10 to 20 times more destructive than bare-foot traffic.
Studies by Liddle & Greig-Smith (1975b) have shown that wear by vehicles and
human trampling on sand dunes reduces species diversity as well as population size
and distribution. They conclude that the consequences of this type of impact are as
great as the major natural environmental factors.
The majority of research on the effect of human trampling on coastal dunes has
been directed toward the response of the vegetation (i.e. specific species)
immediately exposed to the impacts. This type of impact research has also been the
major focus in other ecosystems (LaPage, 1967, 1968; Goldsmith et al., 1970;
Willard & Marr, 1970, 1971; Dale & Weaver, 1974; Foin et al., 1977). In
contrast, a second approach which examines the distribution and species
composition of plant communities which exist in areas which have had long-term
trampling pressures and adjacent areas with little or no trampling pressure provides
an assay of the effect of long-term trampling on vegetation patterns.
The objective of this study is to determine the effect of long-term trampling
pressure on the composition and horizontal pattern of coastal dune vegetation in the
Parker River National Wildlife Refuge.
THE STUDY AREA
The site selected was the dune system on Plum Island, Essex County, Massachusetts
(Fig. 1). The island and its flora were described by McDonnell (1979a,b). Two-thirds
of this 12.8km-long by approximately 0.6km-wide barrier island has remained
relatively natural. About 1,230 ha of the southern portion of the island are now
included in the Parker River National Wildlife Refuge and a Commonwealth of
Massachusetts State Park, which owns the extreme southern tip. This is one of the
largest semi-natural barrier beach dune systems in New England north of Cape Cod.
Plum Island formed after the retreat of the Wisconsin Ice Sheet some 6,300 years
ago (Mclntire & Morgan, 1964). It is composed primarily of unconsolidated beach
and dune sands and marsh deposits (Jones & Cameron, 1976). The southern two-
thirds of the island exhibits classical features of a barrier beach dune system. Four
TRAMPLING EFFECTS ON COASTAL DUNE VEGETATION 29
~er
~ sland
N i ~ '--~ RJ.
co Atlantic Ocean
(¢
Plum Island ,o
Sound
0
Z
"7
State P a r k -~-.--d metres
0 300 900
Fig. 1. A m a p of Plum Island showing the location of the study areas in the Parker River Nationa
Wildlife Refuge, Insert is a m a p of New England showing the location of Plum Island.
292 M.J. McDONNELL
major physiographic zones can be identified running east to west across the island
(Fig. 2): (1)beach; (2)foredune; (3)interdune; and (4)backdune.
The adjacent ocean is an important moderating factor in the climate. The average
annual temperature for this region is 10 °C with a January average of - 1.4 °C and a
July average of 19.9 °C. Average annual precipitation is 193.1 cm with the winter
season being the wettest due to frequent storms. Winds are primarily westerly or
offshore. The storm winds on the other hand are most prevalent from the east, with
those most important to beach formation coming from the northeasterly direction.
BEACH FOREDUNE INTERDUNE = BACKDUNE
L Tr ~ '
Ac
.c ~qt~
~'I
Ps
~'-X RC
Ab LI . Ht ~ , ~
A.
Co
Pm Ac Ps
Ss Mp /
B.
Fig. 2. A diagrammatic sketch of the typical physiographic features and vegetation zones encountered
on two east to west transects across Plum Island: A, area of the refuge with little or no trampling pressure;
B, area exposed to heavy trampling pressure. (Key: Ab, Ammophila breviligulata: Ac, Amelanchier
canadensis: Co, Cehis occidentalis: Cp, Carex pennsylvanica: Ht, Hudsonia tomentosa; Lj, Lathyrus
japonicus; Mp, Myrica pennsylvanica; Pm, Prunus maritima; Ps, Prunus serotina: Rc, Rosa carolina;
Ss, Solidago sempervirens; and Tr. Toxicodendron radicans).
Because of the proximity of Plum Island to a large metropolitan area (Boston),
the refuge management has carefully regulated the number of vehicles entering the
refuge area, particularly during heavy-use periods. Access to the dunes has been
limited to ten parking areas spaced at intervals along the 9.7kin road to the
southern end of the island. Thus trampling pressure has been concentrated in
specific well-defined areas of the dune system. In addition, to limit trampling on the
dune vegetation adjacent to these parking areas, boardwalks to the beach were
established over the past six years. The refuge management has also set aside a large
Dune Natural Area for research which receives very little trampling pressure. Since
the late 1960s the refuge management has not allowed vehicles on the dunes; thus
T R A M P L I N G EFFECTS O N C O A S T A L D U N E V E G E T A T I O N 293
any trampling is caused by either humans or wildlife (e.g. dogs, rabbits, etc.). Plant
nomenclature follows Fernald (1950). Voucher specimens of plants were collected
and deposited in the Hogdon Herbarium, University of Hampshire (NHA).
METHODS
A preliminary examination of the impact of human trampling on the dune
vegetation in the Parker River National Wildlife Refuge revealed that the foredune
and interdune areas were the most severely impacted. These areas are primarily
covered with perennial herbs, e.g. Ammophila breviligulata and Lathyrusjaponicus.
while the backdune vegetation is dominated by woody trees and shrubs. A t'e~.
distinct paths extend through the tree and shrub colonies surrounding each parking
area but it is not until the path reaches the open interdune and foredune areas that
large-scale destruction and even elimination of vegetation can be observed.
To obtain quantitative information on the horizontal structure of the vegetation
in heavily trampled and relatively untrampled foredune and interdune areas, eight
study areas were established. Five study areas were adjacent to parking lots.
alternating among the ten which provide access across the dunes. Study areas were
established adjacent to parking lots 1 (P 1), 3 (P3), 5 (P5), 8 (P8) and 10(P10) (Fig. l).
The study areas were then grouped according to trampling intensity. Accurate
estimates of the actual number of visitors or cars present at a parking lot over a
season are not available. A request by the author to distribute questionnaires to
determine parking lot use patterns was denied due to refuge management policy.
Visual observations of each parking lot over an extended time were beyond the
resources of this project. However, it is felt that, because the emphasis of the study is
on long-term trampling impacts on the area, current parking lot use patterns
obtained from first-hand observations and information from refuge personnel.
along with knowledge of the history and size of each parking lot+ provide an
adequate estimate of trampling intensity. Thus, estimates of trampling intensity for
each parking lot study area were based on (l)current parking lot use trends~
(2) capacity of parking lot to holds cars; (3) number of years 9pen; (4) distance trom
refuge entrance; and (5) length of time boardwalks were established.
Over the past ten years there has been a dramatic increase in the number of people
using the refuge. During a single year, from October 1976 to September 1977. ~ total
of 469,482 people visited the refuge (Parker River National Wildlife Refuge
records). The greatest number occurs during the summer months with an average of
over 50,000 people per month. With only one entrance to the refuge, at the northern
end of Plum Island, there is a distinct gradient of recreational use down the island.
The closer the parking lot is to the entrance the greater use it gets. Thus, parking lots
]-6 get many more visitors than parkings lots 7 11. In general most people do not
drive 9.6 km down a dirt road to the parking lots at the southern end of the island.
From May through August when the weather is pleasant, all of the parking lots
294 M.J. McDONNELL
are filled close to capacity. In the last few years they also fill up rapidly during the
autumn months ( P R N W R records). With such heavy use the capacity of the parking
lots provides a good estimate of the intensity of trampling on the adjacent study
area. The smaller the parking lot, the less intense the trampling and vice versa.
Thus, an area of dune adjacent to a small parking lot at the southern end of the
refuge is less intensely trampled than an area adjacent to a large parking lot at the
northern end of the refuge.
Using this information, coupled with the number of years a parking lot has been
open and the length of time it has had a boardwalk, the parking lot study areas were
grouped into three categories according to level of probable trampling pressures:
severe P1; heavy P3, P5; and moderate P8, PI0 (Table 1).
TABLE 1
PARKING LOT STUDY AREAS GROUPED INTO THREE CATEGORIES OF PROBABLE TRAMPLING PRESSURE
ACCORDING TO AGE, SIZE, DISTANCE FROM REFUGE ENTRANCE AND NUMBER OF YEARS BOARDWALKS WERE
PRESENT
Approximate Distance Total years Estimated
Study Parking Capacity no. of years from refuge boardwalk trampling
area lot (no. of cars) open entrance present pressure
P1 1 145 30 0.09 k m 2 Severe
P3 3 31 28 0.70 k m 6 Heavy
P5 5 22 24 2.08 k m 6 Heavy
P8 8 9 13 6.40 k m 3 Moderate
P 10 10 26 13 8-80 k m 3 Moderate
In addition, three study areas (N1, N2 and N3) were established in relatively
undisturbed areas as far from potential trampling pressure as possible (Fig. 1). Two
study areas were established in the Dune Natural Area: N1 approximately 1200 m
south of parking lot 5; N2 approximately 1600 m north of parking lot 6. Study area
N3 was established in a relatively undisturbed and inaccessible area of the dunes
between parking lots 8 and 9.
Each study area was 100 x 100m with the edges oriented north-south and
east-west. The eastern edge of a study area was the foredune ridge. The study areas
adjacent to parking lots were set up so that the midpoint of the eastern boundary was
the intersection of the eastern end of the boardwalk and the foredune ridge. Those
study areas not adjacent to parking lots were established so that the midpoint of the
eastern boundary was a point originally established on a map and then located in the
field. In each study area four 10 x 100 m transects running north south were centred
on 10, 35, 65 and 95 m respectively from the eastern edge of the foredune ridge. Each
transect was further divided into 40 5 x 5 m quadrats, situated as pairs on either
side of a centre line. Due to the fragile nature of the dune system and refuge
management policy, destructive biomass sampling was not attempted. Instead,
percent cover and frequency values were obtained to determine species composition
and abundance of the vegetation in the study areas.
T R A M P L I N G EFFECTS O N C O A S T A L D U N E V E G E T A T I O N 295
During August 1978 20 randomly chosen 5 x 5 m quadrats in each transect were
sampled for percent cover of species present, total number of species present, and
total area with vegetation cover. Quadrats were chosen at random because the
position of the boardwalks shifts from year to year within about a 50-m area. Thus,
no attempt was made to quantify the gradient of trampling away from the
boardwalks since the exact intensity of trampling at any one point has varied over
the years depending on the position of the boardwalks. The data were statistically
analysed using analysis of variance techniques. Because the quadrats sampled were
randomly chosen they could not be pooled to form the experimental error term for
they may not have the same variance as postulated in the mathematical model. To
provide a more rigorous test the percent cover values were averaged over transects,
and an error term was constructed using selected single degree freedom sums of
squares for transect and location contrasts, set up prior to data analysis. The exact
procedure used follows Cochran & Cox (1957) and is described in McDonnell
(1979a). After completion of the analysis of variance, treatment means were
compared using Least Significant Difference (LSD) at alpha 0-05.
RESULTS A N D D I S C U S S I O N
Twenty-five species were present in the eight study areas. Ammophila breviligulata,
Lathyrus japonicus, Artemisia stellariana and Hudsonia tomentosa together
account for over 75 ~o of the total area covered with vegetation in the study areas
(Table 2). The percent cover of the less dominant species is presented in Table 3.
TABLE 2
THE PERCENT COVER OF THE DOMINANT SPECIES AND TOTAL AREA COVERED WITH VEGETATION AND THE
A V E R A G E NUMBER OF SPECIES PRESENT IN EACH S T U D Y AREA
L S D , L e a s t Significant Difference. In the a n a l y s i s o f v a r i a n c e t h e r e were f o u r t r a n s e c t s a n d eight
l o c a t i o n s ; in n o case w a s the t r a n s e c t X l o c a t i o n i n t e r a c t i o n significant, F-test significant at 0.05 level.
Species Stud)' areas
PI P3 P5 P8 PIO NI N2 N3 ANOVA +
Trampling Average LSD Transect Location
pressure: severe heavy moderate undisturbed
Ammophila
breviligulata 6-7 18.0 20.6 30'6 22.5 18-0 18-8 25.0 20"0 8.5 NS °
Lathyrus
japonicus 2.8 9-9 2.4 12.4 3.2 9.0 4-9 9.0 6.7 6.1 " NS
A rtemisia
stellariana 0" 1 2.0 1-4 2.6 1 "2 7-6 1 "2 0"0 2.0 9.7 NS NS
Hudsonia
tomentosa 0.0 0.5 0.3 0-2 5-3 3-0 8.6 9.0 3.3 6-1 ° "
Total
v e g e t a t i o n cover 9.6 36.0 25.8 46.0 39.0 40-0 43.9 53.4 38.4 14-2 NS "
Average
n o . o f species 0.6 2.0 2.5 2.7 2-8 3.2 3.4 4.0 2.7 0.46 "
N S = n o t significant. " = significant.
296 M . J . McDONNELL
TABLE 3
THE PERCENT COVEROF THE LESSDOMINANTSPECIESWHICH OCCUR IN THE EIGHTSTUDYAREASLOCATEDIN
THE PARKER RIVER NATIONALWILDLIFEREFUGE
Trampling pressure: PI, severe; P3 + P5, heavy; P8 + P10, moderate; NI, N2 and N3, undisturbed.
Stud), areas
Species PI P3 P5 P8 P10 N1 N2 N3
Solidago sempervirens 0"00 0.31 0.41 0.80 1.03 2.34 1.51 0.65
Euphorbiapolygonifolia 0"02 0.03 0.43 0.03 0.18 0.10 0.28 0.08
Cyperus sp. 0.00 0"00 0.00 0.00 0.07 0.13 0.11 1.14
Polygonella articulata 0"00 0"02 0.00 0.00 0.06 0.10 0.14 0.76
Cakile edentula 0.00 0.04 0.10 0.38 0.03 0.00 0-03 0-02
Rosa sp. 0-00 0.87 0-00 0.00 0.13 0.00 3-00 0.00
Prunus serotina 0"00 0.00 0.00 0"00 0.06 0"00 0.00 0.00
Polygonum scandens 0.00 0.00 0-00 0-00 0-01 0-00 0-00 0.00
Spiraea tomentosa 0.00 0.00 0.00 0"00 0"00 0.00 0.43 0.00
Viburnum recognitum 0'00 0'00 0"00 0"00 0"00 0"00 0-43 0'00
Fraxinus pennsylvanica 0-00 0.00 0.00 0.00 0.00 0-00 0.25 0"00
Xanthium eehinatum 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00
Toxicodendron radicans 0.00 4.18 0.00 0.00 1.62 0.00 2.81 0-00
Myrica pennsylvanica 0-00 0-80 0-00 0-00 4.00 0-00 1-43 0-00
Raphanus raphanistrum 0-00 0'00 0-00 0.18 0"00 0"00 0-00 0"00
Prunus maritima 0.00 0.00 0.00 0.00 0.00 0.26 0.04 0.00
Lechea maritima 0"00 0"00 0"00 0"00 O.13 0"00 0"00 O"18
Spartina patens 0.01 0"00 0-00 0-00 0.00 0.00 0.00 0.00
Distichlis spicata 0.01 0'00 0"00 0"00 0'00 0'00 0.00 10.00
Cladonia sp. 0.00 0.00 0.00 0.00 0.44 0"00 0"00 4.50
Geaster sp. 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.03
An analysis of variance (Table 4) of the percent cover of the total area covered
with vegetation revealed that differences among transects (variation within study
areas) were not significant, whereas differences between locations (different
trampling pressures) were. The parking lot study areas as a group have significantly
less vegetation cover than the undisturbed study areas. When the means for each
study area are compared (Table 2) only parking lot study areas PI and P5 (with
severe and heavy trampling pressure respectively) exhibited significantly less
vegetation cover than the undisturbed study areas (N1, N2 and N3). Moderately
trampled parking lot study areas P8 and P10 had significantly more vegetation than
study areas P1 and P5 but were not significantly different from the undisturbed
study areas or study area P3 (heavy trampling pressure). The relatively high cover
value for study area P3 results partially from the fact the boardwalk was moved
50 m north of its previous location just before the study began, thus excluding some
of the more heavily impacted area.
The most important species stabilising the foredune and interdune zones is
A. breviligulata. The analysis of variance (Table 4) indicates that, like total
vegetation cover, the only significant factor was location (i.e. different trampling
pressures). But unlike the total vegetation cover analysis, the undisturbed compared
with disturbed study areas contrast was not significant. Only severely trampled
study area P1 had significantly less percent cover of A. breviligulata than the
undisturbed study areas (Table 2). Study areas with heavy and moderate trampling
TRAMPLING EFFECTS ON COASTAL DUNE VEGETATION 29"]
TABLE 4
A N A L Y SI S OF V A R I A N C E TABLE F O R THE TOTAL AREA COVERED WITH VEGETATION AND Ammophila
breviligulata
Four transects were sampled at each of the eight study areas. N, undisturbed study areas, N l, N2, N 3; P.
parking lot study areas, PI has severe trampling pressure, P3 and P5 have heavy trampling pressure, and
Pg and PI0 have moderate trampling pressure.
Total vegetation cover Ammophila breviligulata
Sums Mean Sums Mean
Sources of variation DF of squares square oj squares square
Overall mean 1 872796 258898
Transect (T) 3 6484 2161 798 266
Location (L) 7 103880 14840 ~ 26322 3760"
LI (N vs. P) 1 32374 32374 ~ 128 128"
L2 (N1 vs. N2) 1 354 354 12 12
L3 (NI + N2 vs. N3) l 6468 6468 2170 2170
L4 (P8 vs. PI0) 1 1904 1904 2621 2621
L5 (P3 vs. P5) 1 4270 4270 228 228
L6 (P8 + P10 vs. P3 + P5) 1 10981 10981" 4039 4039
L7 (P1 vs. P3 + P5 + P8 + PI0) 1 47517 47517 ~ 17125 17125"
L x T 6 7914 1319 7896 1316
Experimental error 15 26378 1758 14184 945
" Significant: F-test, alpha = 0.05.
pressure (P3, P5 and P8, P10, respectively) did not have significantly less cover of
A. breviligulata (Table 2). In fact, study area P8, with moderate trampling pressure,
had significantly higher percent cover of A. breviligulata than undisturbed study
areas N 1 and N2. In the trampled parking lot study areas A. breriligulata made up
50',~;~ 80 ~'0 of the total vegetation cover while in the undisturbed study areas it
o, o/
contributed only 42/o 47/,, to the total cover. It appears that long-term moderate
trampling pressure does not affect total cover, but favours A. hreviligulata over
other more sensitive species.
An analysis of variance of the percent cover of the other dominant species.
L.japonicus, A. stellariana and H. tomentosa, revealed that only the last could he
used to estimate the effect of trampling. Location (different trampling pressures) had
no significant effect on the total cover of L.japonicus or A.,stellariana (Table 2).
Both of these species occur sporadically throughout the dune system.
An AN OVA table for the percent cover of H. tomentosa (Table 5) indicates that,
unlike the preceding species, both transect and location factors were significant. The
percent cover of H. tomentosa was significantly greater in transects C and D, which
are farthest from the foredune, than A and B, which are on the foredune. This species
commonly forms large mats behind the foredune in the stable protected areas of the
interdune (Fig. 2). When the percent cover means for each study area are compared
(Table 2), the undisturbed study areas exhibit significantly more H. tomenlosa than
any of the trampled study areas except PI 0. which has moderate trampling pressure.
No H. tomentosa was present in the severely trampled study area P1, while in heavily
and moderately trampled study areas P3, P5 and P8 respectively, it never occurred
in amounts greater than 1 '~, cover.
298 M.J. McDONNELL
TABLE 5
ANALYSIS OF VARIANCE TABLE FOR THE PERCENT COVER OF nudsonia tomentosa AND FOR THE AVERAGE
NUMBER OF SPECIES PRESENT IN THE EIGHT STUDY AREAS
Hudsonia tomentosa Average number o.] species
Sums Mean Sums Mean
Source o]variation DF of squares square DF of squares square
Overall m e a n 1 7388 1 4665
Transect(T) 3 5788 192 a 3 129 434
L o c a t i o n (L) 9 8249 1178 a 7 576 82 a
T × L 9 4955 550 6 11 2
Experimental error 12 5802 483 15 98 6
Significant; F-test, a l p h a = 0.05.
The increased importance of A. breviligulata and the virtual elimination of
H. tomentosa in areas of the dune exposed to trampling pressure suggest that
trampling has caused the typical foredune community dominated by
A. breviligulata to extend inland, reducing the width of the interdune community.
Thus, the interdune community has become restricted to a narrow strip near the base
of the backdune ridge (Fig. 2).
In general, plant species diversity in the foredune and interdune regions of the
dune system is low. An analysis of variance on the average number of species present
in the study areas revealed that both transect and location factors were significant
(Table 5). Transects farthest from the foredune (C&D) exhibited the greatest
species diversity. A comparison of the average number of species in each study area
(Table 2) revealed that a significantly greater number of species occurred in the
undisturbed study areas as opposed to the study areas exposed to trampling
pressure (average 3.5 spp. area in the undisturbed sites compared with 2.3 spp. area
in areas with heavy trampling pressure). Thus, long-term trampling pressure, in
addition to changing the horizontal pattern of the coastal dune vegetation, has also
reduced species diversity.
CONCLUSION
The effects of long-term human trampling on the coastal dune vegetation in the
Parker River National Wildlife Refuge varied depending on the intensity of
trampling pressure. Severe trampling pressure caused a drastic reduction in both
vegetation cover and species diversity (Fig. 3). Somewhat less conclusive, heavy
trampling pressure appeared to cause a reduction in cover and diversity. Moderate
trampling pressure significantly reduced species diversity, but not total cover. Long-
term moderate trampling appears to have favoured A. breviligulata over other more
sensitive species. In contrast, all levels of trampling pressure significantly reduced
the amount of H. tomentosa present. Even though moderate long-term trampling
pressure has not caused significant and obvious changes in vegetation cover, similar
T R A M P L I N G EFFECTS ON C O A S T A L D U N E V E G E T A T I O N 299
80' "4
70 Average <~
Number of S p e c i e s \
\
w
\
\
60 \ 3
\
a
\
\
50 \.
w
\
,ota, \
"4'
Vegetation Cover ~~ /~
\ 2 w
40 \
\ \
\
\
O
30
\
m
J
AmmoDhila [ ~
Z
20 brevilioulata
\ • <
10
Hudsonia
tomentosa
I I
UNDISTURBED MODERATE HEAVY SEVERE
N1, N2, N3 P8, P10 P3, P5 Pt
TRAMPLING PRESSURE
Fig. 3. A summary of the effects of different trampling pressures on the coastal dune vegetation in the
Parker River National Wildlife Refuge. N and P refer to numbered study areas. Solid symbols indicate a
significant difference from undisturbed study areas at alpha = 0.05.
to severe and heavy trampling, it has caused a loss o f species, a shift in relative
a b u n d a n c e o f species, and a shift in the horizontal pattern o f the vegetation. The
increased importance o f A. breviligulata, and the virtual elimination o f interdune
species such as H. tomentosa in all areas exposed to trampling pressure, indicates
that long-term h u m a n trampling has resulted in the expansion o f the foredune
c o m m u n i t y d o m i n a t e d by A. breviligulata and the restriction o f the interdune
c o m m u n i t y to a narrow strip at the base o f the backdune ridge (Fig. 2).
300 M . J . McDONNELL
Smaller p a r k i n g lots a n d the e s t a b l i s h m e n t of boardwalks has successfully
reduced the impact of t r a m p l i n g on the vegetation in the foredune a n d i n t e r d u n e
regions in the Parker River N a t i o n a l Wildlife Refuge. Results from this study
suggest that destruction of d u n e vegetation by t r a m p l i n g can be minimised if d u n e
access is restricted to specific areas with p a r k i n g lots designed to hold only 10 to 20
cars a n d have b o a r d w a l k s that extend over the d u n e s to the beach.
ACKNOWLEDGEMENTS
I would like to express my gratitude to Dr G a r r e t t E. Crow and D r Russell K i n e r s o n
for their help t h r o u g h o u t this study. I wish to t h a n k George Gavutis, m a n a g e r of the
Parker River N a t i o n a l Wildlife Refuge: w i t h o u t his c o o p e r a t i o n this study could n o t
have been completed. The following persons assisted with the field work: Lisa
B a n d a z i a n , L y n n Bohs, Allison Buck, Bill C o t r e a u a n d Dennis M c K e a r i n . Dr Jerry
W a r r e n a n d the staff of the office of Academic C o m p u t i n g at U N H provided
valuable assistance in the statistical analysis of the data. I would also like to t h a n k
Dr Richard T. T. F o r m a n and an a n o n y m o u s reviewer for their constructive criticism
of the m a n u s c r i p t . This study was supported in part by two Central University
Research G r a n t s ( C U R F $86 a n d S106) from the University of New Hampshire.
REFERENCES
BLOM, C. W. P. M. (1977). Effect of trampling on soil compaction on the occurrence of some Plantago
species in coastal sand dunes, II. Trampling and seedlingestablishment. Oecol. Plant., 12, 363 81.
BOORMAN, L. A. & FULLER, M. (1977). Studies on the impact of paths on the dune vegetation at
Winterton, Norfolk, England. Biol. Conserr., 12, 203 16.
BROADHEAD,J. M. & GODFREY,P. J. (1977). Off road vehicle impact in Cape Cod National Seashore:
Disruption and recovery of dune vegetation. Int. J. Bioclim. Biomet., 21,299 306.
BURDEN.R. F. & RANDERSON,P. F. (1972). Quantitative studies of the effects of human trampling on
vegetation as an aid to the management of semi-natural areas. J. appl. Ecol., 9, 439 58.
CLARK, J. (1974). Coastal ecosystems. Washington, DC, Conservation Foundation.
COCHRAN,W. & Cox, G. (1957). Experimental design. New York, John Wiley.
DALE, D. & WEAVER,T. (1974). Trampling effects on vegetation of the trail corridors of north Rocky
Mountain forests. J. appl. Ecol., 11, 767-72.
FERNALD,M. L. (1950). Gray's manual oJbotany, 8th edn, corrected edn, 1970. New York, Van Nostrand
Reinhold.
FOIN,T. C., GARTON,E. O., BOWEN,C. W., EVERINGHAM, M., SCHULTZ,R. O. & HOLTON,G. (1977).
J.
Quantitative studies of visitor impacts on environments of Yosemite National Park, California, and
their implications for park management policy, J. environ, mgmt, 5, 1 22.
GOLDSMITH,F. B., MUNTON,R. J. C. & WARREN,A. (1970). The impact of recreation on the ecologyand
amenity of semi-natural areas: methods of investigation used in the Isles of Scilly. Bin. J. Linn. Soc.,
2, 287 306.
JONES,J. R. & CAMERON,B. (1976). Sedimentaryand geomorphicorigin and developmentof Plum Isla.nd,
Massachusetts: An example of a barrier island system. In Geology oJSouthern New England. A
guidebook jor field trips to the Boston area and vicinity, ed. by B. Cameron, 188 204. Boston, Science
Press.
TRAMPLING EFFECTS ON COASTAL DUNE VEGETATION 301
LAMOUREUX,G. & GRANDTNER,M. M. (1977). Contribution/~ l'etude ecologique des dunes mobiles, 1
Les 61~ments phytosociologique. Can. J. Bot., 55, 158 71.
LAPAGE, W. P. (1967). Some observations on campground trampling and ground cover response. U.S
Forest Service Res. Paper N. E. 68.
LAPAGE, W. P. (1968). Recreation and the forest site. J. For., 66, 319 21.
LIDDLE, M. J. (1975a). A selective review of the ecological effects of human trampling on natural
ecosystems. Biol. Conserv,, 7, 17-36.
L1DDLE, M. J. (1975b). A theoretical relationship between the primary productivity of vegetation and its
ability to tolerate trampling. Biol. Conserv., 8, 251 5.
LIDDLE, M. J. & GREIG-SMITH, P. (1975a). A survey of tracks and paths in a sand dune ecosystem. 1. Soils.
J. appl. Ecol., 12, 893 908.
LIDDLE, M. J. & GREIG-SMITH, P. (1975b). A survey of tracks and paths in a sand dune ecosystem. 11
Vegetation. J. appl. Ecol., 12, 909 30.
MARTIN, W. E. (1959). The vegetation of Island Beach State Park, New Jersey. Ecol. Monogr., 29.1 46
McDONNELL, M. J. (1979a). The vascular flora oJ Plum Island. Essex Count.l', Massachus'ett.s. with an
analysis of the impact oJhuman trampling on coastal dune regetation. MS thesis, University of New
Hampshire, Durham, NH.
MCDONNELL, M. J. (1979b), The flora of Plum Island, Essex County. Massachusetts. Bull. New Hamp~
agric. Exp. Stn, 513.
MCINTYRE, W. & MORGAN, J. (1964). Recent geomorphic history oj Plum Island, Massachu.sett.s. and
adjacent coasts. Baton Rouge, Louisiana State University Press.
NICKERSON, N. (1976). Effects of pedestrian traffic on sand dune vegetation. In Abstract,s ./rom
ConJerence on Scientific Research in the National Parks. 1st. New Orleans, Louisiana. American
Institute of Biological Sciences and National Parks Service. United States Department of the
Interior.
N1EDORODA,A. & LIMEBURNER,R. (1975). A preliminary report on the geomorphic effects of recreational
vehicles on coastal systems. National Park Service Cooperative Park Studies Unit Rept., No 7
Amherst, Ma.. University of Massachusetts.
OLDFIELD, F., CARTER,R., KITCHER,K. & WILCOCK. F. (1972). Report oj investigations into coastal
erosion and accretion along the coastlines oj County Antrim and County Londonderry. ('oasta/
Research Project Report 196~1972. School of Biological and Environmental Studies. New
University of Ulster.
OLSEN, S. B. & GRANT, M. J. (1975). Rhode Island's Barrier Beaches: Volume 1. A report on a
management problem and an evaluation of options. A Coastal Management Publication
University of Rhode Island. Marine Technical Report No.4. Kingston. R.I.
OOSTING, H. J. & BILLINGS, W. D. (1942). Factors effecting vegetational zonation on coastal dunes.
Ecology, 23, 131-42.
RICHARDS, 1. & STEAD, F. (1978). Dune conservation in Anglesey. Tn Ctrv Plann., 46, 173 6.
TREW, M. J. (1973). The effects and management of trampling on coastal sand dunes, d. Environ. Plann. &
Pollut. Control. 1, 38-49.
VAN DER VALK,m. G. (1974). Mineral cycling in coastal foredune plant communities in Cape Hatteras
National Seashore. Ecology, 55, 1349 58.
VAN DER VALK, A. G. (1975). The floristic composition and structure of foredune plant communities of
Cape Hatteras National Seashore. Chesapeake Sci., 16, 127 33.
WILLARD, B. & MARR, J. (1970). Effects of human activities on alpine tundra ecosystems in Rock 5
Mountain National Park, Colorado. Biol. Conserv., 2, 257-65.
WILLARD, B & MARR, J. (1971). Recovery of alpine tundra under protection after damage by human
activity in the Rocky Mountains of Colorado. Biol. Conserv., 3, 131 90.
WILLIS, A. J., FOLKES, B. F., HOPE-SIMPSON,J. F. & YEMM, E. W. (1959a). Braunton Burrows: The dune
system and its vegetation, Part I. J. Ecol.. 47, 1 22.
WILLIS, A. J.. FOLKES,B. F., HOPE-SIMPSON,J. F. & YEMM, E. W. (1959b). Braunton Burrows: The dune
system and its vegetation, Part II. J. Ecol.. 47, 249 87.