Ecological Studies of the Sunken Forest,
Fire Island National Seashore, New York

NPS Scientific Monograph No. 7
NPS Logo


To gain a wider understanding of the ecological relationships of the Sunken Forest, an analysis of the vegetation between the ocean and bay was undertaken. Although the main emphasis was the vegetational analysis of the maritime forest community, plant communities leeward and windward of the forest were also investigated since all these communities are interrelated. Martin (1959) found that despite the apparent highly regular distribution of species and communities on barrier islands, very few species are restricted to a single topographic facet. Due to the close, interrelationships between geological, biological, and meteorological processes on barrier islands, the investigation of the Sunken Forest was conducted in the context of the entire range of plant communities on Fire Island.

Previous studies of the Sunken Forest area by Murphy (1933), Thorne (1953), and Schulte (1965) have described the vegetation in general, nonquantitative terms. The vegetational analysis here was necessary to provide quantitative information, such as species density, basal area, cover, and frequency, necessary for the study of biogeochemical relationships in the Sunken Forest. The salt marsh, dune, and swale communities were included in the vegetational analysis to provide an understanding of the ecological processes of the entire barrier island and the interactions between the maritime forest and the communities surrounding it.


Vegetation in the Sunken Forest area was sampled in a 570-m-wide zone from the Atlantic Ocean to the Great South Bay (Fig. 14). The vegetation was divided into three communities: (1) the dune and swale area; (2) the Sunken Forest proper; and (3) the salt marshes. Delineations of community types on barrier islands have been based largely on plant growth forms: herbaceous, shrubby, and arborescent (Oosting 1954; Boyce 1954; Davis 1957). However, the interspersion of shrubby and herbaceous vegetation in the dune area reduced the utility of a growth-form approach in defining the communities in the Sunken Forest area.

Fig. 14. Vegetation map of the Sunken Forest area. Dune and swale transects are marked A, B, C. Forest sampling plots are small squares, while rectangle marked P is the ecosystem analysis plot. (click on image for an enlargement in a new window)

The dune and swale community consisted of the vegetation between the winter storm high water line on the beach and the forest community on and behind the secondary dune. The Sunken Forest community extended northward from the boundary of the dune and swale community and terminated with the disappearance of vegetation dominated by trees. The salt-marsh communities occupied the area between the forest and the bay and were dominated by herb or shrub species. Each of these communities was sampled between June and August 1967.

The Dune and Swale Community

Three randomly selected, meter-wide belt transects were laid out from the mean tide mark to the beginning of the forest community (Fig. 14). The transects were comprised of a series of 1 X 1-m plots. In each plot percentages were estimated for: (1) cover, by species; (2) bare sand; and (3) litter visible without disturbing the vegetation. A topographic profile of each transect was constructed by the use of an Abney level and a meter tape.

Transects across the dunes and swale gave evidence of spatial changes in the community but did not accurately estimate the vegetation cover of the community as a whole (Table 1). Therefore, a vegetation map (Fig. 14) was constructed by projecting an aerial photograph of the Sunken Forest area (Fig. 15) onto a sheet of paper and outlining vegetational boundaries; the patterning of the dune and swale vegetation being characterized by discrete patches of vegetation composed of several species and dominated by a single species. These patches of vegetation are termed "species-type units." The original map had a scale of 1:1430, large enough for field identification of these species-type units in the dune and swale community.

aerial photo of Sunken Forest
Fig. 15. Aerial photograph of the Sunken Forest area taken 15 March 1966. True North is at top of the photograph. This plate is a copy of the original aerial photograph (scale 1:4800) used in the vegetational analysis. (Courtesy National Park Service and Brookhaven National Laboratory).

Table 1. Dune and swale vegetation cover.

Transect A Transect B Transect C Averages
% cover % cover % cover % cover % of total cover

Arctostaphylos uva-ursi23.115.710.816.641.5
Ammophila breviligulata8.
Prunus maritima2.
Hudsonia tomentosa1.
Myrica pensylvanica3.
Smilax glauca3.
Parthenocissus quinquefolia3.10.5
Solidago sempervirens1.
Ilex opaca0.11.5
Lathyrus japonicus0.1
Rhus radicans0.
Cirsium horridulum0.
Aralia nudicaulis0.6**0.20.6
Quercus stellata0.7

Lachea maritima0.
Pinus rigida0.20.4
Panicum spp.
Artemisia stelleriana0.1
Quercus velutina0.3
Rhus copallina0.2**0.10.3
Smilacina stellata0.1
Agrostis alba
Cakile edentuta**0.1*0.1
Carex pensylvanica0.1***0.1
Hieracium venosum0.1
Sassafras albidum
Amelanchier canadensis*

Artemisia caudata
Carex silicea**
Cyperus rotundus
Euphorbia polygonufolia
Lactuca canadensis
Polygonella articulata
Pteridium aquilinum
Rosa carolina*

Smilax rotundufolia

Solidago odora*

Trientalis borealis0.1


* = <0.1%

The species-type units were carefully cut from the map and weighed. From the map area:paper weight ratio, the land surface areas covered by the various species-type units were estimated. The percent cover of the dominant species within the various species-type units was estimated in randomly selected units in the dune and swale community. The average cover of dominant species for the entire dune-swale community was calculated by multiplying the surface areas of the species-type units by the actual species cover within the units.

The Sunken Forest Community

A sampling grid representing 10 X 10-m squares was laid over the aerial photo of the Sunken Forest and 34, 10 X 10-m plots were randomly chosen to sample the forest vegetation.

These plots were located in the field by pacing along compass bearings from known landmarks. All individuals with a diameter greater than 3.0 cm at breast height (1.5 m) were tallied as trees by species. Heights were estimated or measured to the nearest 0.1 m and diameters were measured to the nearest millimeter. If the stem of an individual plant branched below breast height, each of the branching trunks was tallied as a separate individual. Diameters at breast height (dbh) of standing and fallen dead trees were counted by species.

The 10 X 10-m plot was subdivided into four 5 X 5-m plots and one was randomly selected for sampling the shrub layer. Shrubs, individuals greater than 1 m high and with a dbh of less than 3.0 cm, were tallied by species. Heights were measured to the nearest 0.1 m and basal diameters at the ground surface were measured to the nearest millimeter. Each shrub shoot emerging from the soil was tallied as a separate individual.

The herbaceous layer was sampled in 1 X 1-m plots laid out in each of the four corners of the 10 X 10-m plot. Any plant under 1 m in height was considered to be an herb. Percent cover estimates were made for each species present, except for Smilax species, for which the densities in stems/m2 were measured.

Phenological records were made during trips to the Sunken Forest which occurred at roughly weekly intervals between October 1968 and October 1969. During each visit to the forest, the condition of the vegetation, stages of flowering, and leaf condition were noted.

The phenology of litter fall was analyzed from collections of litter in 12 baskets situated in an ecosystem analysis plot in the Sunken Forest (Fig. 14). These baskets, with a collecting area of 1780 cm2 each, were anchored to stationary positions throughout the study, and emptied at approximately weekly intervals. The material falling into the baskets was separated by species (Ilex, Sassafras, Amelanchier, and others) and then by components (leaves, flowers plus fruit, and twigs plus wood). All litter components were dried at 85°C and weighed.

The Salt Marsh Community

A meter-wide belt transect was laid out through the center of one of the salt marshes on the southern edge of the forest. The 33-m transect ran from the edge of the forest to the Great South Bay. Within the transect, percent species cover was estimated in a manner similar to the analysis of the dune and swale community.

<<< Previous <<< Contents >>> Next >>>

Last Updated: 21-Oct-2005