| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ecology |
Halophyte Biotechnology Center, College of Marine Studies, University of Delaware, 700 Pilottown Road, Lewes, Delaware 19958 USA
Received for publication October 8, 2002. Accepted for publication February 18, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: American beachgrass • Ammophila breviligulata • Assateague Island • dune morphology • feral horses • grazing • Poaceae • sand dune • Spartina patens
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
).
The ponies graze on the plants growing on the sand dunes and in the salt marshes of the island (Furbish and Albano, 1994
). Ammophila breviligulata Fern., American beachgrass, the dominant grass vegetating the primary dunes, is an important component of the wild horses' diet (Bashore and Furbish, 1991
). Spartina patens, salt meadow hay, occupies some dunes as well and is also grazed but is much less frequent in occurrence. The horses visit the foredunes to graze, especially during the summer months, when the breezy oceanside of the island offers a cooler temperature with fewer insects and biting flies. Because the horse population has increased from no more than 20 in the 1960s to approximately 170 today, there is concern that overgrazing in the dune and marsh habitats may hasten erosion and alter the ecology of the island. Ammophila breviligulata, with its extensive root system, is very important in stabilizing the sand dunes and thus controlling erosion. In addition, A. breviligulata initiates dune formation by stimulating sand deposition and accumulation.
The goal of this research was to assess the effects of the horses' grazing on the dominant grasses on the island's foredunes, primarily A. breviligulata and secondarily Spartina patens. Spartina patens was dominant on only a few dunes. The findings of this 2-yr study will aid the National Park Service in their population management program for the feral horses to assure sustainable horse and natural dune plant populations.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). However, within this zone, for purposes of this study, four types of foredune formations were distinguished, all of which were included in the study (Figs. 1–4). They were (1) sand flats and areas blown out or overwashed during storms, i.e., no dune present; (2) small, rounded, and newly forming dunes forming knolls; (3) long, continuous dune ridges; and (4) butte-shaped dunes scarped by erosion. Throughout this paper, these will be referred to as flats, knolls, ridges, and buttes, respectively. The dominant vegetation on all four dune types was most commonly Ammophila breviligulata and secondarily Spartina patens.
|
Because of the different shapes of these dunes and the area covered by beachgrass, the dimensions of the paired plots varied with dune type. Both the fenced and unfenced plots in the flat sites were 15 x 20 m in size. The knoll and butte plots were 13 x 13 m, and the ridge plots were 8 x 24 m in size. Within the 15 x 20 m plots, 12 subplots were delineated at the time of sampling, nine subplots within the 13 x 13 m plots, and 12 subplots within the 8 x 24 m plots. Comparisons of growth and cover of A. breviligulata or S. patens between fenced and unfenced plots were made for each dune type separately.
Field sampling
During May and September 1994 and September 1995, percentage cover, percentage flowering, and plant height data were collected within the fenced and unfenced plots. The subplots within the plots were marked temporarily using wire flags each time we sampled, and then the flags were removed immediately after the data were collected. To mark the subplots, wire flags were placed at every 4 or 5 m, depending on plot dimensions. Thus, there were 12 25-m2 (5 x 5 m) subplots in each 15 x 20 m plot, nine 16-m2 (4 x 4 m) subplots in each 13 x 13 m plot, and 12 16-m2 (4 x 4 m) subplots in the 8 x 24 m plots. Data sheets consisted of one sheet per plot, divided into the appropriately sized subplots for each dune type. Data for each subplot was recorded in each box, and the location of the vegetation was marked on each subplot map.
Using the Daubenmire cover scale (Mueller-Dombois and Ellenberg, 1974
), the estimated percentage plant cover was determined for each species for each subplot within the fenced or unfenced plot, thereby sampling the entire plot, rather than random samples. Although A. breviligulata and S. patens were the predominant species on these dunes, cover was recorded for all species in each subplot. Cover classes were as follows: 0–1%, 1–5%, 5–25%, 25–50%, 50–75%, 75–95%, 95–100%. A plant stand was considered to be 100% covered if, when looking at it from one end of the subplot, very little sand surface was visible because of the dense vegetation. In natural dunes in the mid-Atlantic region, 100% cover is rarely seen, but in managed fertilized dunes it is not unusual.
Percentage flowering data (the percentage of the total number of stems that were in flower) were estimated and recorded for each subplot. Flowering was determined in late summer/early fall. Plant height of ungrazed flowering and nonflowering stems, along with inflorescence length in the case of A. breviligulata, were measured on the plants collected in the 0.1-m2 biomass plots described below. Plant height data were also collected nondestructively in the field by randomly selecting and measuring the height of six mature ungrazed nonflowering and flowering stems per fenced or unfenced plot. The inflorescence length of the six flowering plants (A. breviligulata) was also measured.
Aboveground harvest data
In September 1995, in addition to the same nondestructive collection of data described, we harvested plant material to determine aboveground biomass, stem density, leaf length and width, and stem diameter. To collect the additional 1995 harvest data, we selected one site in each whole fenced or unfenced plot in which we placed a 0.1-m2 quadrat. Because the vegetation density is so diverse in the dunes, even within a single plot, we harvested a quadrat of the densest vegetation in that particular plot to ensure that the sample was indicative of the maximum biomass and stem density for a given plot. The percentage cover of the 0.1-m2 quadrats was estimated, and then plants were cut at the sand surface. Samples were bagged and stored in a refrigerator until they could be measured, dried, and weighed. Biomass samples were dried in a 60°C oven to constant mass. Biomass estimates for the entire plot were calculated based on the dry mass and the percentage cover of the 0.1-m2 quadrat and the percentage cover of each subplot.
Belowground harvest data
After the aboveground biomass was sampled, all the belowground biomass was dug from the 0.1-m2 quadrat. In the laboratory, roots, rhizomes, and belowground shoot material were washed, separated, and dried in a 60°C oven to constant mass. Belowground shoots were measured and counted.
Data analysis
Data were analyzed within dune types with one-way ANOVA using SPSS software (SPSS, Chicago, Illinois, USA). Because one of the S. patens plot pairs converted from S. patens to primarily A. breviligulata during the study, it was omitted from the analyses. Unfortunately, this left only two replications for this comparison.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Percentage cover
Percentage cover at the end of the growing season in 1995 is shown in Tables 1–4 for A. breviligulata, and the change in cover over the period from May 1994 to September 1995 is shown in Table 5. Except for the ridge plots, cover was significantly less in the unfenced plots than in the fenced plots (Tables 1– 4). Table 5 indicates the amount by which the percentage cover changed between May 1994 and September 1995. During the 1994 growing season, cover increased significantly in the fenced plots, but either did not change or exhibited only a slight increase in the unfenced plots. While cover in the fenced plots continued to increase between September 1994 and September 1995, cover in the unfenced plots decreased (flat, knoll, and butte plots) or remained the same (ridge plots).
|
|
|
|
Plant morphology
Plant height in both fenced and unfenced plots was measured on nongrazed stems only. Height of plants in the fenced plots of all dune types was about the same, ranging from 50.2 to 52.6 cm (Tables 1–4). A significant difference in plant height between fenced and unfenced plots was seen in the flat plots (P = 0.001), where there was a difference of 15 cm, and in the butte plots (P = 0.076), where there was a difference of 11.5 cm.
Except for the ridge plots, stem density was greater in the fenced than in the unfenced plots. Leaf length was significantly greater in the flat and knoll fenced plots. Leaf width did not vary significantly with fencing. Stem diameter was greater in the fenced butte and ridge plots.
Aboveground biomass
Biomass was dependent on the presence or absence of fencing. Except for the ridge plots, biomass was significantly greater in the fenced than in the unfenced plots (Tables 1–4). Biomass also varied considerably with dune type. In the fenced plots, the least biomass was found in the flats at 112 g/m2, followed by the knolls at 224 g/m2, then by the ridge sites at 315 g/m2, and finally by the buttes at 346 g/m2.
Belowground biomass
In Tables 1–4, belowground biomass is divided into two components, roots + rhizomes and buried shoots. In fenced plots, root and rhizome biomass was significantly greater than that in the unfenced plots only in the flat plots. With buried shoot biomass, however, this difference was only significant in the butte plots. The amount of sand burial was significantly greater in the fenced (vs. unfenced) flat and butte plots.
Reproductive effort
Three characteristics related to flowering were significantly different. First, we found a much higher percentage of plants in flower in the fenced vs. the unfenced plots (Table 6). This also held true for the ridge plots which, as seen before, did not exhibit differences in many of the other characteristics measured. Interestingly, the percentage flowering in the fenced ridge, knoll, and butte plots was about the same, 64.6–69.4%, while in the fenced flat sites, the percentage flowering was significantly lower (<30%). Second, the height of the ungrazed flowering stems in the fenced plots was significantly greater than that of flowering plants in the unfenced plots of three of the four dune types—flat, ridge, and butte. Third, the length of the inflorescence also differed significantly between plants in fenced vs. unfenced plots with smaller inflorescences typical of the unfenced plants. This measurement was of course taken only on ungrazed stems.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Differences in percentage plant cover between fenced and unfenced plots were great in the flat, knoll, and butte plots (Tables 1, 2, and 4), while fenced and unfenced ridge plots exhibited almost no significant differences (Table 3). The ridge plots may have been less affected because grazing at these sites is less intense than at the other sites. The horses spend less time at the southern end of the park than they do further north (C. E. Furbish, Assateague Island National Seashore, personal communication), and the ridge plots are all located in the more southern portion of the park.
|
|
found A. breviligulata biomass to increase with increasing dune age on Hog Island, Virginia, USA. Biomass was of course also dependent on the presence or absence of fencing. Except for the ridge plots, biomass was significantly greater in the fenced than in the unfenced plots (Tables 1–4). Once again, the fenced and unfenced ridge plots appear to be more similar, possibly from the less intensive grazing in the southern part of the park. In the flat sites, vegetation is just coming in after severe storms have washed through and flattened the dunes. The plants' ability to recover in these newly established areas may be less following grazing. Because they have fewer resources belowground, grazing puts a greater stress on these plants, thus making the differences between the flat and the other dune types more obvious. In general, there was more loss in 1995 than in 1994, probably because the grass weakened in 1994 had fewer reserves with which to build a new canopy in 1995. It is obvious from Fig. 5 that the flat type is the most vulnerable to the detrimental impacts of grazing. In the particular plot shown, cover dropped to 0–3% in just two growing seasons and would likely disappear completely the following year at this location. Vegetation on the other dune types is less vulnerable in the short term. Ammophila breviligulata reproduces both vegetatively and sexually. Studying the change in percentage cover is a good means of evaluating localized vegetative reproduction. Equally important is the effect of grazing on sexual reproduction because reproduction by seed is an important means of propagule dispersal. The percentage of seed germination for A. breviligulata is high (unpublished data), and seeds allow for dispersal by wind and water. Interestingly, the percentage flowering in the fenced ridge, knoll, and butte plots was about the same, approximately 65–69%, while in the fenced flat sites the percentage flowering was significantly lower (<30%). This may reflect the younger age of the plant stands in these newly establishing sites and their limited belowground resources. The shorter height of the ungrazed flowering stems in the unfenced plots likely affects the distance seeds would travel via wind dispersal, and the smaller inflorescence size typical of these stems indicates less seed production by plants exposed to grazing.
These data demonstrate a decrease in propagation potential even when a particular stem was not grazed that season, or at least not grazed late in the season. The differences seen in all three characteristics measured (Table 6) may be due to lower carbohydrate reserves stored in previously grazed vs. ungrazed "parent" plants. If aboveground vegetation is reduced by grazing, a reduced amount is available for photosynthesis and therefore for storage in the rhizomes. Therefore, in the spring when new growth begins, less reserves are available to send up new and vigorous shoots. There are fewer reserves for flower production as well, and because the new shoots are fewer and less vigorous than those of ungrazed plants, there is less aboveground material to photosynthesize and create new fuel during that season for flower and seed production. Therefore, flowers are fewer and smaller, reducing the number of seeds and perhaps the quantity and quality of the carbohydrate reserves in the seeds necessary for successful germination and seedling growth.
Spartina patens was dominant on only a few dunes, newly establishing flats, in contrast to the dunes of Hog Island further south where S. patens produced more biomass than A. breviligulata on all but the oldest dunes (Dilustro and Day, 1997). Although we had only two plot pairs to analyze, we did find significant differences between fenced and unfenced plots, indicating that horse activity significantly affected this species as well.
The other species in our plots, Solidago sempervirens, Panicum amarulum, Myrica pensylvanica, and Cenchrus tribuloides, were not affected by the exclosures. However, numerous other grazing studies in the sand dunes (Zeevalking and Fresco, 1977
; Hewett, 1985
; Gibson, 1988
; Hester et al., 1994
; Kooijman and Smit, 2001
) and salt marshes (Jensen, 1985
; Bazely and Jefferies, 1986
; Belanger and Bedard, 1994
; Furbish and Albano, 1994
; Olff et al., 1997
) have demonstrated significant grazing effects on species composition. In fact, grazing is used in Europe as a management tool to maintain plant diversity in the dunes (Hewett, 1985
; Boorman, 1989
; Kooijman and Smit, 2001
). That we did not find grazing effects on species composition in this study is likely due to the extremely dynamic nature of the foredunes of Assateague Island. Only four subplots of the 342 total number of subplots in the 16 plot pairs analyzed had vegetation cover over 50%. Exclosures increased cover significantly, but rarely above 50%. The dunes are naturally sparsely vegetated. Therefore, regardless of the presence or absence of grazing, competition is not significant enough to inhibit other species from coming into the area. A longer term study might eventually show otherwise, but I think this would more likely be possible for the less dynamic secondary dune area rather than the foredune area of this study. Observations of foredune vegetation along the Delaware, Maryland, and Virginia coasts where grazing does not occur indicate sparse vegetation as well. In fact, the state of Delaware conducts an extensive annual spring planting of A. breviligulata to stabilize dunes that eroded during the previous winter's storms. The dense vegetation typical of many European dunes is very uncommon on the very active dunes along the mid-Atlantic and southeastern Atlantic coasts of the United States. The dynamic nature of the mid-Atlantic dunes actually helps maintain the health of A. breviligulata, and when the dynamics are reduced, vigor declines (Eldred and Maun, 1982
; Disraeli, 1984
; Maun and Lapierre, 1984
; Seliskar, 1994
, 1995
).
In conclusion, grazing by the number of horses currently present on Assateague is detrimental to the growth and spread of A. breviligulata. Grazing by the horses reduces plant cover, vegetative spread of the plants, biomass, flowering, and seed production. Therefore, without the fencing and with the current number of horses, vegetation density will continually decrease, making the dunes more vulnerable to erosion. Initially, I thought that grazing by the numerous deer inhabiting the island may cause significant damage as well. However, there was sometimes evidence of grazing in the fenced plots, i.e., deer and/or rabbit pellets present near grazed stems. Although the fencing may have deterred most deer, deer and rabbits can enter and exit the fenced plots. Despite this, the vegetation in the fenced plots increased significantly. Vegetation is critical for stabilizing the sand dunes, and without the extensive root and rhizome systems of A. breviligulata and S. patens to hold the sand in place, the dunes are very vulnerable to erosion. Not only is the root system important, but the shoots also trap sand by decreasing the velocity of the wind, thus causing the sand to drop near the plants, thereby initiating dune formation. In this study, sand accumulation was significantly greater in the ungrazed than in the grazed flat and butte plots. This accumulation correlated with significantly greater plant height and stem density in the ungrazed plots, which may have enhanced sand deposition by decreasing wind velocity.
Reducing the size of the horse herd would likely be very beneficial to the stabilization of the dunes along Assateague's coast. A birth control program has been initiated by the National Park Service. By gradually decreasing the number of horses, determining a herd size at which the horses and the vegetation are in balance should be possible. In addition, small patches of dunes could be fenced to serve as propagule sources for both vegetative and sexual reproduction. And, if the beach is nourished by sand replenishment in some sections, planting A. breviligulata along the upper beach zone while restricting horse access during plant establishment would allow the grass to initiate dune formation.
|
FOOTNOTES |
|---|
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Bazely D. R. R. L. Jefferies 1986 Changes in the composition and standing crop of salt-marsh communities in response to the removal of a grazer. Journal of Ecology 74: 693-706[CrossRef]
Belanger L. J. Bedard 1994 Role of ice scouring and goose grubbing in marsh plant dynamics. Journal of Ecology 82: 437-445[CrossRef]
Boorman L. A. 1989 The grazing of British sand dune vegetation. Proceedings of the Royal Society of Edinburgh 96B: 75-88
Carter R. W. G. K. F. Nordstrom N. P. Psuty 1990 The study of coastal dunes. In K. F. Nordstrom, N. Psuty, and B. Carter [eds.], Coastal dunes: form and process, 1–14. Wiley, New York, New York, USA
Dilustro J. J. F. P. Day 1997 Aboveground biomass and net primary production along a Virginia barrier island dune chronosequence. American Midland Naturalist 137: 27-38[CrossRef][ISI]
Disraeli D. J. 1984 The effect of sand deposits on the growth and morphology of Ammophila breviligulata. Journal of Ecology 72: 145-154[CrossRef]
Eldred R. A. M. A. Maun 1982 A multivariate approach to the problem of decline in vigor of Ammophila. Canadian Journal of Botany 60: 1371-1380[ISI]
Furbish C. E. M. Albano 1994 Selective herbivory and plant community structure in a mid-Atlantic salt marsh. Ecology 75: 1015-1022[CrossRef][ISI]
Gibson D. J. 1988 The relationship of sheep grazing and soil heterogeneity to plant spatial patterns in dune grassland. Journal of Ecology 76: 233-252[CrossRef]
Hester M. W. B. J. Wilsey I. A. Mendelssohn 1994 Grazing of Panicum amarum in a Louisiana barrier island dune plant community: management implications for dune restoration projects. Ocean and Coastal Management 23: 213-224[CrossRef]
Hewett D. G. 1985 Grazing and mowing as management tools on dunes. Vegetatio 62: 441-447[CrossRef][ISI]
Jensen A. 1985 The effect of cattle and sheep grazing on salt-marsh vegetation at Skallingen, Denmark. Vegetatio 60: 37-48[CrossRef][ISI]
Keiper R. R. 1985 The Assateague ponies. Tidewater Publishers, Centreville, Maryland, USA
Kooijman A. M. A. Smit 2001 Grazing as a measure to reduce nutrient availability and plant productivity in acid dune grasslands and pine forests in The Netherlands. Ecological Engineering 17: 63-77[CrossRef][ISI]
Maun M. A. J. Lapierre 1984 The effects of burial by sand on Ammophila breviligulata. Journal of Ecology 72: 827-839[CrossRef]
Mueller-Dombois D. H. Ellenberg 1974 Aims and methods of vegetation ecology. John Wiley & Sons, New York, New York, USA
Olff H. J. De Leeuw J. P. Bakker R. J. Platerink H. J. Van Wijnen W. De Munck 1997 Vegetation succession and herbivory in a salt marsh: changes induced by sea level rise and silt deposition along an elevational gradient. Journal of Ecology 85: 700-814
Seliskar D. M. 1994 The effect of accelerated sand accretion on growth, carbohydrate reserves, and ethylene production in Ammophila breviligulata (Poaceae). American Journal of Botany 81: 536-541[CrossRef][ISI]
Seliskar D. M. 1995 Coastal dune restoration: a strategy for alleviating dieout of Ammophila breviligulata. Restoration Ecology 3: 54-60[CrossRef][ISI]
Zeevalking H. J. L. F. M. Fresco 1977 Rabbit grazing and species diversity in a dune area. Vegetatio 35: 193-196[CrossRef][ISI]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
