EFFECTS OF HUMAN DISTURBANCE ON THE DUNE VEGETATION
         OF THE GEORGIA SEA ISLANDS



                   John C. Rodgers, III
                 Department of Geosciences
                 Mississippi State University
                    P.O. Drawer 5448
             Mississippi State, Mississippi 39762-5448
                 rodgers@geosci.msstate.edu


      Abstract: This research investigates how human disturbance has affected foredune veg-
    etation of the Georgia Sea Islands (GSI) in the United States. The cover of native dune-
    building grasses (Uniola paniculata L. and Panicum amarum Ell.) was more abundant
    within less-disturbed sites than in sites that had higher levels of human disturbance. In
    contrast, dunes in human-disturbed areas had significantly higher cover of alien plants and
    native generalist taxa, and they also had higher overall species diversity. Additionally, the
    cover of native dune-building grasses was significantly greater on protected National
    Wildlife Refuge islands than on more frequently visited and developed tourist islands. In
    addition to the ANOVA, nonmetric multidimensional scaling (NMS) analysis showed that
    the vegetation composition differed between disturbed and less-disturbed plots and
    between plots on tourist islands and protected islands. Both ANOVA and NMS analyzes
    agree that dunes in human-modified areas have lower dune-grass cover and greater cover
    of species that are not adapted to building and stabilizing dunes. Therefore, human distur-
    bance may indirectly reduce dune stability by altering the dune vegetation. [Key words:
    dune vegetation, human disturbance, Georgia Sea Islands.]



                     INTRODUCTION

  Coastal dune communities are developed and maintained through a complex
interaction of vegetation colonization and sand accumulation. Initially, beach sand
is colonized by low-growing plants termed “dune initiators” that grow by spreading
laterally across the sand. The vegetative bodies of the dune initiators act as a sub-
strate for coastal dune formation because they trap and protect seeds and other
propagules of the subsequent colonizers, the perennial rhizomatous grasses (Wood-
house, 1982). When these perennial grasses become established, they intercept
wind-borne sand with their tussocks and trap it with their extensive root/rhizome
system (Wagner, 1964), which causes dunes to grow. In this way, these grasses act
as “dune builders” and they are a crucial factor for coastal dune formation (Wood-
house, 1982; Heyligers, 1985).
  Dune formation is a positive feedback cycle where the accumulation of sand
stimulates the growth of dune grasses, which in turn results in the further accumu-
lation of sand (Woodhouse, 1982). Previous work has shown that dune morphology
is linked to the type of vegetation that colonizes the dunes (Heyligers, 1985; Nord-
strom, 1990). On the Georgia coast, for example, dunes that develop under tall

                          79
Physical Geography, 2002, 23, 1, pp. 79–94.
Copyright © 2002 by V. H. Winston & Son, Inc. All rights reserved.
80                JOHN C. RODGERS, III



grass species obtain larger size and steeper dip angles (>25%) than dunes that are
colonized by smaller grasses and forbs (Oertel and Larson, 1976).
  Because coastal dune morphology and vegetation are strongly linked, any alter-
ation of the species composition or cover can have profound effects on the dune
size and shape. Human disturbance (defined here as human activities that have
resulted in changes in community and ecosystem structure and composition
through the alteration of the physical environment; Drake et al., 1989; Pyle, 1995)
is a factor that has been known to affect dune vegetation. For example, introduction
of the European beachgrass (Ammophila arenaria L.) in southeastern Australia has
transformed the morphology of many coastal dunes (Heyligers, 1985). Dunes that
were once small (<2 m) and had gentle slopes have become larger (up to 6 m) with
steeper slopes (>30%) within only 20 years as a result of colonization by alien spe-
cies. The arrival of A. arenaria in Australia also has caused new dunes to form in
areas where they had not previously occurred (Heyligers, 1985). Similarly along the
Pacific United States, dunes have formed along the Oregon coast where they have
been historically absent due to the introduction of A. arenaria and A. breviligulata
Fern. (Cooper, 1958; Wiedemann and Pickart, 1996) and dunes in California have
changed from being low and parabolic to becoming vertical and immobile from the
introduction of A. arenaria (McPherson, 1994).
  Besides introduction of species, humans also have physically damaged the veg-
etation cover of dunes in many coastal areas. For example, trampling by foot and by
vehicles directly destroys native dune vegetation and alters dune soil properties
(e.g., soil moisture), which leads to lower overall dune stability (Rickard et al.,
1994; Rust and Illenberger, 1996). Furthermore, dunes with lower native species
cover become unstable and eventually deflate because many of these species can-
not maintain themselves on eroding sand surfaces (Wagner, 1964). Therefore,
human modification of dunes, by providing habitats for alien species, by changing
the dune environment, and/or by disrupting the accretion feedback cycle, can have
profound influences on coastal dune ecosystems.
  The objective of this research is to investigate the role of human disturbance on
the dune vegetation of the Georgia Sea Islands (GSI). The specific questions I
address are: (1) Does the cover of native dune builders, the cover of alien species,
the species diversity, and the openness of vegetation differ between dune sites that
have a high degree of human disturbance and sites that have a lower degree of
human disturbance? (2) Do broad-scale, island-level human use patterns (tourist
islands versus protected islands) affect local site vegetation cover? By addressing
these research questions, I provide insight into how human disturbance can affect
the southeastern United States coastal dune ecosystems.

                  STUDY AREA

  The GSI (Fig. 1) have varying degrees of human use, but for this study I classified
islands as either tourist islands or protected islands depending on the overall land
use and accessibility from the mainland. I sampled the vegetation cover from two
tourist islands and from two protected islands. Even though the location of these
islands varied in latitude, the island flora was rather uniform across the entire study
         EFFECTS OF HUMAN DISTURBANCE ON DUNE VEGETATION          81




 Fig. 1. Georgia Sea Islands used in this study.


region because north-south climatic gradients on the Georgia coast are muted and
because each island had essentially the same set of potential colonists (Stallins,
2001). Compositional differences on these islands, then, are probably related to
other factors, like human disturbance history.
  The two tourist islands were Jekyll Island and St. Simons Island. Jekyll Island is
managed as a state park through a board of trustees, the Jekyll Island Authority.
Within the developed areas, Jekyll Island has many hotels, residential areas, and
businesses. The island is accessible by vehicle via a causeway, and in 1997 Jekyll
Island had 1.5 million visitors (Jekyll Island Convention and Visitors Bureau, pers.
comm., April 1999). The island is approximately 20.4 km2 in area and 16 km in
length. St. Simons was the other tourist island used in this study. Like Jekyll Island,
St. Simons is connected to the mainland via a causeway and it has many hotels and
residential areas. It was the largest island (80.3 km2 in area and 17 km long) and it
was perhaps the most heavily developed island used in this study. But even though
it was heavily developed, there existed pockets of less-disturbed dunes. During
1997, over 1.7 million people visited St. Simons Island (Brunswick and The Golden
Isles of Georgia Visitors Bureau, pers. comm., April 1999).
82                JOHN C. RODGERS, III



  The two protected islands were Blackbeard and Wassaw Islands. Blackbeard
Island is a National Wildlife Refuge that is managed through the U.S. Fish and Wild-
life Service (USFW). It is approximately 11 km long and has an area of 22.7 km2.
The island is one of the oldest National Wildlife Refuges and has been protected
since 1924. Although the island is open to the public, it is only accessible by boat
and transportation is not provided by USFW. According to the Savanna Coastal Ref-
uge USFW office, very few visitors come to the island because of its inaccessibility
(USFW, pers. comm., October 2002). Except for a few residents who stay on the
island to monitor sea turtle nesting and for a few USFW officers, Blackbeard Island
is uninhabited. Wassaw Island also is a National Wildlife Refuge that is only acces-
sible by boat. Until recently, the entire island was privately owned, but in 1976 the
owners donated most of the land to the USFW. Today a small part of the island is
still privately owned by the original family, but most of the island is uninhabited.
Wassaw Island is approximately 10 km long and has an area of 43.5 km2. Like
Blackbeard Island, the Wassaw Refuge is open to the public but few visitors come
to the island because of its inaccessibility. However, both Blackbeard and Wassaw
Islands have deer hunts for a few weeks during the fall and winter months when 50
to 100 hunters camp on the islands.


                   METHOD

Field Sampling

  I used the point intercept method (Goodall, 1957) to sample the absolute cover
of all herbaceous and woody taxa (<1.5 m in height) on the foredunes of Black-
beard, Jekyll, St. Simons, and Wassaw Islands. On each island, I had a total of 10
dune sites; 5 sites were located in areas that had a high degree of human distur-
bance (hereafter referred to as disturbed sites) and 5 sites were located in areas that
had a lower degree of human disturbance (hereafter referred to less-disturbed sites).
Aerial photographs, orthophotoquads, and ground reconnaissance were used to
identify potential sites before sampling. The criteria for accepting sampling sites
were that disturbed sites occurred within 1 m of heavily used boardwalks; hiking
trails and vehicle access roads that cut through dunes (which existed on both the
protected and tourist islands); dunes with high foot traffic; or, in the case of a few
sites on Blackbeard Island, dunes that had been mechanically smashed to foster the
nesting of sea turtles. Less-disturbed dune sites were at least 30 m away from dis-
turbed areas. Even though not as common, less-disturbed dunes existed on both
tourist islands. Owing to within-island longitudinal differences in sediment accre-
tion and erosion on the GSI (Hayes, 1994), all sites were located on the southern
half of the islands where dunes were more prominent. Vegetation sampling was
conducted in 1997 during September and October, when the dune vegetation is at
its most mature stage (Wilbur Duncan, University of Georgia, Athens, pers. comm.,
September 1997).
        EFFECTS OF HUMAN DISTURBANCE ON DUNE VEGETATION           83


Data Analysis

  In the following analyses, key dune species were placed into several vegetation
classes. First, the principal native dune-building species in this study were the
perennial, rhizomatous grasses Uniola paniculata L. and Panicum amarum Ell.,
both of which are commonly responsible for the development and stabilization of
dunes in the southeastern United States (Woodhouse, 1982). I investigated differ-
ences in cover of this vegetation class as a whole and for each individual dune-
building species. Second, alien species included those species that were described
as either “nonnative,” “alien,” “exotic,” or “introduced.” Alien status was deter-
mined from descriptions in Small (1913), Godfrey and Wooten (1979, 1981), Hitch-
cock (1950), Sanders (1987), and U.S. Department of Agriculture (1971), and the
species encountered in this study that met these descriptions were Chloris patrea
Swartz., Cynodon dactylon (L.) Pers., Digitaria sanguinalis (L.) Scop., Paspalum
notatum Flugge., Sorghum halepense (L.) Pers., Lantana camara L., and Lonicera
japonica Thunb. Not all of these species were present on all the islands (see the
Results section). Analysis of this group was based on the combined cover values of
all the alien species present at a particular site. Third, generalist species were those
native species that occur in many different inland habitats besides coastal dunes
(Castillo and Moreno-Cassasola, 1996). This vegetation class consisted of Rubus L.
spp., Smilax L. spp., Cenchrus tribuloides L., and Hydrocotyle bonariensis Comm.
These species were placed into a single group for discussion purposes, but the sta-
tistical analyzes were performed on the individual taxon because they were each
commonly found within the study area.
  Differences in cover were examined for certain individual species. As mentioned
earlier, U. paniculata L. and P. amarum Ell. were tested individually to investigate
the extent to which the cover of each dune-building grass varied. I examined differ-
ences in cover of one particular alien species, C. dactylon, because it was the most
frequent and dominant dune alien species on all islands. I also examined differ-
ences in cover for native dune taxa that were common (frequency = 50%) or that
had high cover values (cover = 30%) in at least five sites. These important species
included Spartina patens (ait.) Muhl, Heterotheca subaxillaris (Lam.) Britt. and
Rusby, Iva imbricata Walt., Ipomoea stolonifera (Cyr.) Gmel., and Croton punctatus
Jacq. Besides these vegetation classes and individual species, I calculated species
diversity for each site using the Shannon-Weiner Index (Barbour et al., 1999) and
the absolute cover of bare ground (area where no species were recorded) for a
proxy, but inverse relationship with vegetation density.
  Two-factor analysis of variance (ANOVA; Sokal and Rohlf, 1981) was used to
investigate among site differences in dune-building species cover, alien species
cover, species diversity, open ground, and cover of prominent taxa. Differences in
vegetation were examined between disturbed and less-disturbed sites and among
islands. The research hypothesis tested by these analyses was that vegetation pat-
terns between tourist islands and protected islands and between disturbed and less-
disturbed sites would be significantly different with regards to the vegetation
classes, individual species, and species diversity. The absolute cover values were
transformed using the arcsine of the square root (Zar, 1999) to normalize the data.
84                     JOHN C. RODGERS, III



   Table 1. ANOVA of the Effects of Island and Disturbance Category on the
   Absolute Cover of Native Dune Builders, Alien Species, Species Diversity,
              and Bare Ground Per Island

               Native dune builders  Alien species  Species diversity  Bare ground
Source of variation        (Pr > F)       (Pr > F)     (Pr > F)     (Pr > F)

Island               .093*        .257       .38       .291
Disturbance             .0001**       .037**      .02**      .503
Interaction             .155         .303       .39       .07*


* Significant at p < .10.
** Significant at p < .05.


  Nonmetric multidimensional scaling (NMS) was used to investigate overall veg-
etation differences among sites. NMS is an ordination technique that has been
widely adopted by ecologists and is used to organize and make sense of complex
compositional data sets (Kenkel and Orióci, 1986; Minchin, 1987; Clarke, 1993).
NMS graphically positions sites in a multidimensional ordination space based on
their species composition. Distances between sites represent compositional similar-
ity (Kenkel and Orióci, 1986), such that sites positioned closer to one another share
many of the same species in similar abundances while far away sites are more dis-
similar. As an aid to interpreting the NMS results, I used Varimax rotation to col-
lapse the multiple dimensions into a two-dimensional ordination in which each site
was given specific X, Y coordinates (Clarke, 1993; Legendre and Legendre, 1998).
With these coordinates, I calculated the mean center (average X, Y position) and
corresponding 95% confidence interval for each disturbance category and for each
island. These descriptive statistics are illustrated on the ordinations as ellipses in
which the center of the ellipse represents the mean center, the polar axis represents
the confidence interval around the average Y position, and the equatorial axis rep-
resents the confidence interval around the average X position. The arrangement of
these ellipses helped indicate differences in composition among sites and under-
score ecological responses to human disturbance. PC-ORD Version 4 (McCune and
Mefford, 1999) was the software used to perform NMS ordinations.

                         RESULTS

Spatial Variation of Vegetation Classes and Individual Species

  The ANOVA results showed significant differences between disturbance catego-
ries for the vegetation classes, but differences were not as strong among island
groups (Table 1). Native dune builders were more abundant in less-disturbed sites
(p = .0001). This vegetation class had over six times as much cover in less-disturbed
sites (61% mean cover) as in disturbed sites (11% mean cover) across all islands
(Fig. 2). The ANOVA results also showed significant differences in the cover of
native dune builders among islands (p = .09). Blackbeard and Wassaw Islands had
         EFFECTS OF HUMAN DISTURBANCE ON DUNE VEGETATION                  85




  Fig. 2. Average absolute cover of native dune builders (±SD). Shaded bars represent disturbed sites
and open bars represent less-disturbed sites.




  Fig. 3. Average absolute cover of alien species (±SD). Shaded bars represent disturbed sites and open
bars represent less-disturbed sites.




higher mean cover values of native dune builders (43% mean cover for both
islands) than Jekyll Island (38% mean cover), but all of these had higher cover than
St. Simons Island (12% mean cover).
  Disturbed sites had significantly greater cover of alien species (Table 1). Alien
taxa were found in 14 of the total 40 sites, most of which (85%) were in disturbed
areas. Across all disturbed sites, alien cover accounted for 3.8% of the total mean
cover, whereas mean alien cover in the less-disturbed sites was less than 0.10%
(Fig. 3). Significant differences in alien cover among islands were not consistent. On
both tourist islands, alien taxa were found in all the disturbed sites, and alien spe-
cies were present within 2 nondisturbed sites on St. Simons Island. Alien cover on
the protected islands showed a different pattern. There were 2 sites on Blackbeard
Island that had alien species present (46% cover in a disturbed site and 15% in a
less-disturbed site), but alien plants did not occur in any of the other protected
island sites. Wassaw Island had alien species present within inland areas (Rodgers,
1999), but they were not found in any of the foredune study sites. The inconsistent
variation in alien cover between tourist and protected islands yielded ANOVA
results that failed to reveal any significant differences in alien cover among islands
(p = .257; Table 1).
86                  JOHN C. RODGERS, III




  Fig. 4. Average Shannon-Weiner species diversity index (±SD). Shaded bars represent disturbed sites
and open bars represent less-disturbed sites.




  Fig. 5. Average absolute cover of bare ground (±SD). Shaded bars represent disturbed sites and open
bars represent less-disturbed sites.




  Additionally, disturbed sites had significantly greater species diversity (Fig. 4).
On average, there were three more species in disturbed sites than in less-disturbed
sites. This higher species diversity reflected the greater number of both alien species
and native generalist taxa present within these sites. In terms of among-island vari-
ation, Blackbeard Island, Wassaw Island, and Jekyll Island had greater diversity in
the disturbed sites, but St. Simons Island showed no difference between disturbed
and less-disturbed sites. Diversity on St. Simons Island was equally high in both dis-
turbance categories.
  The cover of bare ground was not significantly different between disturbance cat-
egories (Fig. 5). This indirectly indicates that the average vegetation cover showed
little difference between disturbed and less-disturbed sites across all islands. Varia-
tion in bare ground cover among islands also was not significant; however, the
interaction of disturbance and island was significant (p = .07). The protected islands
had higher bare ground cover in disturbed sites, but the tourist islands, especially
St. Simons Island, had higher bare ground cover in less-disturbed sites (Fig. 5). The
significance of this interaction indicates a possible disturbance-related effect on the
overall dune vegetation cover of protected islands.
          EFFECTS OF HUMAN DISTURBANCE ON DUNE VEGETATION                  87




  Fig. 6. Average absolute cover of frequent dune taxa (±SD) by disturbance category. Shaded bars
represent disturbed sites and open bars represent less-disturbed sites. Asterisks indicate that there is a
significant difference between disturbance categories ( p < .10).



  The response of individual species to disturbance effects was variable. For native
dune-building grasses, Uniola paniculata showed a pronounced difference
between disturbance categories (41% difference in mean cover) whereas Panicum
amarum was not significantly different between disturbance categories (7% differ-
ence in mean cover; Fig. 6). Variations in cover of other important native dune spe-
cies also showed mixed results (Fig. 6). Spartina patens was significantly more
abundant within disturbed sites across all islands (27% mean cover). Croton punc-
tatus, Ipomoea stolonifera, and Iva imbricata had greater cover within the less-dis-
turbed sites, but they were not significantly different among disturbance categories.
Heterotheca subaxillaris showed no difference in cover between disturbance cate-
gories. Generalist taxa were clearly more abundant in disturbed sites across all
islands. In particular, Cenchrus tribuloides, Hydrocotyle bonariensis, Rubus spp.,
and Smilax spp. had significantly greater cover within disturbed sites than within
less-disturbed sites (Fig. 6).
  Most alien species encountered in the study area were Eurasian and tropical
grasses. These included Chloris patrea, Cynodon dactylon, Digitaria sanguinalis,
Paspalum notatum, and Sorghum halepense. All these alien grass species, with the
exception of S. halepense, were found on Blackbeard Island, Jekyll Island, and St.
88                   JOHN C. RODGERS, III




  Fig 7. Average absolute cover of frequent dune taxa (±SD) by island type. Shaded bars represent
tourist islands and open bars represent protected islands. Asterisks indicate that there is a significant
difference between disturbance categories ( p < .10).




Simons Island. S. halepense occurred on both tourist islands, but was not found on
either protected island. The nongrass alien species, Lantana camara, and Lonicera
japonica, were only found on St. Simons Island. C. dactylon was the most common
alien species and had significantly greater cover in disturbed sites across all islands
(Fig. 6).
  Differences among islands in the cover of individual dune species were variable
(Fig. 7). Uniola paniculata and Panicum amarum were not significantly different
among islands. Similarly, most generalist taxa exhibited no significant differences in
cover among islands with the exception of Hydrocotyle bonariensis, which had
extremely high cover values on Wassaw Island (15% mean cover on Wassaw Island
and less than 6% on the other islands). However, the cover of important native dune
species showed definite distinctions among islands. For example, Spartina patens
was substantially more abundant on St. Simons Island, where it accounted for over
60% of the absolute cover in some sites. The other native dune taxa showed clear
differences in cover among protected and tourist islands (Fig. 7). Ipomoea
stolonifera, Iva imbricata, and Heterotheca subaxillaris were more abundant (p =
.001, p = .05, p = .0001, respectively) on the tourist islands while C. punctatus was
more abundant (p = .006) on the protected islands.
          EFFECTS OF HUMAN DISTURBANCE ON DUNE VEGETATION                  89




  Fig. 8. Ordination results from nonmetric multidimensional scaling by disturbance categories. Open
circles indicate less-disturbed sites and shaded circles represent disturbed sites. Ellipses represent the
95% confidence interval around the mean center for disturbed and less-disturbed sites.


Spatial Variation of Georgia Sea Islands Dune Vegetation

  NMS ordination results showed that the overall vegetation composition clearly
differed between disturbance categories (Fig. 8), but island distinctions were not as
strong (Fig. 9). From the ordination, disturbed sites clustered more within the lower
left quadrant while the less-disturbed sites clustered in the upper right quadrant.
Island-based compositional patterns follow a similar trend, but are less distinct (Fig.
9). The tourist islands are fairly well segregated in the upper right quadrant and the
protected islands are bottom and left side of the NMS diagram. This indicates that
the island groups differ in their overall vegetation response to disturbance.

                      DISCUSSION

The Role of Human Disturbance on the Vegetation of the GSI

  Results from this study indicate that dunes in human-disturbed areas had mark-
edly different vegetation than dunes in less-disturbed areas. The ANOVA and NMS
analyses showed that the vegetation composition differed substantially, even
though the overall vegetation cover was similar. Disturbed dunes had reduced
cover of native dune-building grasses and had higher cover of alien species, native
generalist taxa, and other native dune taxa.
  Disturbed dune sites, being in close proximity to boardwalks and vehicle access
points, are places where human traffic is intense. Even on the protected islands,
dune species composition has been affected. Trampling and other forms of human
disturbances directly injure roots and rhizomes of dune grasses and damage the sur-
face sand layers that minimize freshwater evaporation (Rust and Illenberger, 1996).
By destroying the existing native dune vegetation and by disrupting the surface
soils, human disturbance is providing gaps in which generalist, alien, and other
90                  JOHN C. RODGERS, III




  Fig. 9. Ordination results from nonmetric multidimensional scaling by island. Triangles indicated
Blackbeard Island, squares represent Jekyll Island, circles represent St. Simons Island, and diamonds
represent Wassaw Island. Ellipses represent the 95% confidence interval around the mean center for
each island.



native dune taxa can invade from nearby areas. These other species, once estab-
lished, often increase in abundance at the expense of the native dune builders.
  Natural disturbances on barrier islands have been shown to lower species diver-
sity (Ehrenfeld, 1990; Hayden et al., 1995). For example, overwash events on the
North Carolina Outer Banks create hypersaline soils that are mostly inhabited by
Spartina patens (Godfrey, 1976; Woodhouse, 1982). The GSI seldom experience
these natural disturbance agents at this magnitude because they are under the influ-
ence of different wave and tidal regimes (Walker and Coleman, 1987). Where dune
vegetation on the GSI is subjected to human disturbance, dunes have higher plant
species diversity. In that same context, diversity is lower on the GSI in less-disturbed
dune areas because native species tend to dominate. If one ascribes to the interme-
diate disturbance hypothesis (Connell, 1978), these findings suggest that human
disturbance levels on the GSI are operating at centrally located values of intensity.
Dune diversity is high when dunes are subjected to human disturbance and diver-
sity is low when either disturbance is minimal (because native dune species tend to
dominate) or extreme (because only a few species are adapted to harsh conditions).
This is clearly an instance in which maximizing species diversity is not a sound eco-
logical policy.
  Differences in vegetation composition also were apparent between protected
islands and tourist islands, albeit these differences were not as strong. Tourist
islands had lower cover of native dune builders and had higher cover of other
native dune taxa. The degree of trampling and other human modifications are much
greater on the tourist islands because they are visited by more people and they are
more developed. As a consequence, the frequency and magnitude of human distur-
bance are more extensive. Human disturbance is affecting dunes even in the more
sheltered areas on tourist islands. Overall island land-use contrasts are reflected in
species composition at the local site level.
        EFFECTS OF HUMAN DISTURBANCE ON DUNE VEGETATION           91


  The cover of alien species was not significantly different among islands, prima-
rily because of large variances in alien cover values among islands. This was an
interesting finding because it indicates that alien plants were present on all the
islands regardless of the islands accessibility to people. The most common alien
species were Eurasian and tropical grasses that are adapted to human disturbance
and cultivation (Hitchcock, 1950). These species appear to be filling in gaps in the
dunes that were created by human disturbance. It seems, then, that alien plant dis-
tributions on the GSI are more a function of the number of available disturbed sites
on an island than the overall island accessibility or the level of human usage.
  St. Simons was the most heavily disturbed and humanly modified GSI included
in this study, yet the cover of alien species was much less than that of Jekyll or even
protected Blackbeard islands. Several of the heavily disturbed sites on St. Simons
had dunes with lower elevations, which would make them more susceptible to salt-
water inundation. The native generalist and alien taxa that were present on dis-
turbed dunes on other islands are not as common here because they are not able to
tolerate overwash. Indeed, Spartina patens, a well-known overwash colonizer, dis-
played high cover values on several St. Simons Island dunes.

Implications of Human Disturbance and Dune Vegetation

  The replacement of native dune builders with other native dune, alien, or gener-
alist taxa could have profound influences on foredune dune morphology for several
reasons. In contrast to the native dune builders that have an extensive and dense
subsurface root/rhizome system (Woodhouse, 1982), many of the other species
found on disturbed dunes lack adequate root morphology to trap and hold sand.
Additionally, it has been shown that dune height is directly proportional to the cor-
responding height of the dune vegetation (Oertel and Larson, 1976; Heyligers,
1985). The native dune builders on the GSI are efficient at intercepting wind-blown
sand (Wagner, 1964; Woodhouse, 1982), but other species probably do not extend
high enough to contribute meaningfully to sand interception. Cynodon dactylon
has been shown to increase substrate stability in other areas (Mack and D’Antonio,
1998); however, its short stature would make it less capable of intercepting sand.
Lastly, native dune builders have a leaf and stem morphology well adapted to
reducing drag coefficients of winds and causing wind-blown sand to settle out
(Woodhouse, 1982). The other species are not as effective as the native dune-build-
ing grasses at instigating sand deposition. Furthermore, the presence of generalist
and alien taxa may be interfering dune succession. Therefore, dunes in disturbed
areas that have a lower abundance of native dune builders or have a higher cover
of other vegetation taxa are probably be less structurally stable and potentially more
susceptible to erosion than less-disturbed dunes.

                  CONCLUSION

  This research provides evidence that human disturbance has significantly
affected the vegetation composition of the foredune environments on the GSI. At
the local scale, key dune species that trap and bind sand and that are important for
92                   JOHN C. RODGERS, III



the formation and stability of dunes have declined in areas where human distur-
bance has been high. Furthermore, other species that are not as adapted to building
and maintaining dunes, such as other native generalist taxa and alien species, have
increased in abundance on disturbed dunes. Because dune stability and vegetation
cover are inherently linked, this human-induced vegetation change may indirectly
be reducing the stability of coastal dunes. At the broader, island scale, human dis-
turbance on protected islands can lead to localized alterations of foredune vegeta-
tion and environments similar to those experienced on the more heavily used
islands. Likewise, protected dunes on tourist islands maintain less-disturbed rem-
nants of a vegetation composition that are similar to less-disturbed dunes on more
protected islands.

   Acknowledgments: This research was funded by a grant from the National Science Foundation to
Kathleen C. Parker and John C. Rodgers, III (SBR-9701830). Additional funding for this research was
provided by a National Science Foundation grant to Kathleen C. Parker and Albert J. Parker (SBR-
9313704). I am grateful to my doctoral advisor, Kathleen C. Parker, for her guidance and help with the
manuscript. Jenny Cruise, Sam Rodgers, Tony Stallins, and John Schaefer provided assistance in the field.
I also would like to thank Tony Stallins and Wilbur Duncan for assistance with plant identifications, and
for help with the NMS. I am grateful for the Savannah Coastal Refuge USFW and the Jekyll Island Author-
ity for giving me permission to sample on their property.


                      REFERENCES

Barbour, M., Burk, J., Pitts, W., Gilliam, F., and Schwartz, M. (1999) Terrestrial Plant
  Ecology. New York, NY: Addison Wesley Longman.
Castillo, S. A. and Moreno-Cassasola, P. (1996) Coastal sand dune vegetation: An
  extreme case of species invasion. Journal of Coastal Conservation, Vol. 2, 13–22.
Clarke, K. R. (1993) Non-parametric multivariate analysis of changes in community
  structure. Australian Journal of Ecology, Vol. 18, 117–143.
Connell, J. H. (1978) Diversity in tropical rainforest and coral reefs. Science, Vol.
  199, 1304–1310.
Cooper, W. S. (1958) Coastal Sand Dunes of Oregon and Washington Geological
  Society of America Memoir 72. Baltimore, MD: Waverly.
Drake, J. A., Mooney, H. A., DiCastri, F., Groves, R. H., Kruger, F. J., Rejmanek, M.,
  and Williamson, M. (1989) Biological Invasions: A Global Perspective. Chiches-
  ter, UK: John Wiley and Sons.
Ehrenfeld, J. (1990) Dynamics and processes of barrier island vegetation. Aquatic
  Sciences, Vol. 2, 437–480.
Godfrey, P. J. (1976) Comparative ecology of east coast barrier islands: Hydrology,
  vegetation, soils. In Barrier Islands and Beaches: Technical Proceedings of the
  1976 Barrier Island Workshop. Annapolis, MD: The Conservation Foundation, 5–
  34.
Godfrey, R. K. and Wooten, J. K. (1979) Aquatic Plants of the Southeastern United
  States: Monocotyledons. Athens, GA: University of Georgia Press.
Godfrey, R. K. and Wooten, J. K. (1981) Aquatic Plants of the Southeastern United
  States: Dicotyledons. Athens, GA: University of Georgia Press.
        EFFECTS OF HUMAN DISTURBANCE ON DUNE VEGETATION           93


Goodall, D. W. (1957) Some considerations in the use of point quadrat methods for
  the analysis of vegetation. Australian Journal of Biological Science, Vol. 5, 1–41.
Hayden, B. P., Santos, M., Shao, G., and Kochel, R. (1995) Geomorphological con-
  trols on coastal vegetation at the Virginia Coast Reserve. Geomorphology, Vol.
  13, 283–300.
Hayes, M. O. (1994) The Georgia Bight barrier system. In R. A. Davis, ed., Geology
  of Holocene Barrier Islands. Berlin, Germany: Springer-Verlag, 233–304.
Heyligers, P. (1985) The impact of introduced plants on foredune formation in
  southeastern Australia. Proceedings of the Ecological Society of Australia, Vol.
  14, 23–41.
Hitchcock, A. S. (1950) Manual of the Grasses of the United States, Second Edition
  Revised by Agnes Chase. New York, NY: Dover.
Kenkel, N. C. and Orióci, L. (1986) Applying metric and nonmetric multidimen-
  sional scaling to ecological studies: Some new results. Ecology, Vol. 6, 919–928.
Legendre, P. and Legendre, L. (1998) Numerical Ecology. Amsterdam, The Nether-
  lands: Elsevier Science BV.
Mack, M. C. and D’Antonio, C. M. (1998) Impacts of biological invasions on distur-
  bance regimes. Trends in Ecology and Evolution, Vol. 13, 195–198.
McCune, B. and Mefford, M. J. (1999) PC-ORD Multivariate Analysis of Ecological
  Data, Version 4. Gleneden Beach, OR: MjM Software Design.
McPherson, F. (1994) California Native Plant Society Santa Cruz County Chapter.
  Available: http://www.cruzcnps.org/NFduneTxt.html
Minchin, P. R. (1987) An evaluation of the relative robustness of techniques for eco-
  logical ordination. Vegetatio, Vol. 69, 89–107.
Nordstrom, K. F. (1990) The concept of intrinsic value and depositional coastal
  landforms. The Geographic Review, Vol. 80, 68–81.
Oertel, G. and Larson, M. (1976) Developmental sequences in Georgia coastal
  dunes and distributions of dune plants. Bulletin of the Georgia Academy of Sci-
  ence, Vol. 34, 35–48.
Pyle, L. L. (1995) Effects of disturbance on herbaceous exotic plant species on the
  floodplain of the Potomac River. American Midland Naturalist, Vol. 134, 244–
  253.
Rickard, C., McLachlan, A., and Kerley, G. (1994) The effects of vehicular and
  pedestrian traffic on dune vegetation in South Africa. Ocean and Coastal Man-
  agement, Vol. 23, 225–247.
Rodgers, J. C., III. (1999) The Effects of Human Disturbance on Alien Plant Distribu-
  tions and Primary Dune Vegetation on the Georgia Sea Islands. Unpublished
  doctoral dissertation, University of Georgia, Athens.
Rust, I. C. and Illenberger, W. K. (1996) Coastal dunes: Sensitive or not? Landscape
  and Urban Planning, Vol. 34, 165–169.
Sanders, R. (1987) Identity of Lantana depressa and L. ovatifolia (Vergenaceae) of
  Florida and the Bahamas. Systematic Botany, Vol. 12, 44–60.
Small, J. (1913) Flora of the Southeastern United States. New York, NY: Author.
Sokal, J. D. and Rohlf, S. (1981) Biometry, 2nd Edition. New York, NY: W. H. Free-
  man and Co.
94               JOHN C. RODGERS, III



Stallins, J. A. (2001) Soil and vegetation patterns in barrier-island dune environ-
  ments. Physical Geography, Vol. 22, 79–98.
U.S. Department of Agriculture. (1971) Common Weeds of the United States. New
  York, NY: Dover.
Wagner, R. H. (1964) The ecology of Uniola paniculata in the dune strand habitat
  of North Carolina. Ecological Monographs, Vol. 34, 79–125.
Walker, H. J. and Coleman, J. M. (1987) Atlantic and Gulf Coastal Province. In W.
  Graf, ed., Geomorphic Systems of North America. Boulder, CO: Geological
  Society of America, Centennial Special Vol. 2, pp. 51–110.
Wiedemann, A. M. and Pickart, A. (1996) The Ammophila problem on the North-
  west Coasts of North America. Landscape and Urban Planning, Vol. 34, 287–
  299.
Woodhouse, W. W., Jr. (1982) Coastal sand dunes of the U.S. In R. Lewis, ed., Cre-
  ation and Restoration of Coastal Plant Communities. Boca Raton, FL: CRC Press,
  1–44.
Zar, J. (1999) Biostatistical Analysis, 4th Edition. Upper Saddle River, NJ: Prentice
  Hall.