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

NPS Scientific Monograph No. 7
NPS Logo


Results and Discussion

Pattern and process in Fire Island communities

The most striking feature of the Sunken Forest area vegetation, as on barrier islands in general, is the change in species dominance in response to environmental gradients. In addition to the relatively predictable salt spray, sand deposition, wind flow, and cyclic littoral erosion gradients, the complex-gradient on Fire Island includes an erratic and nonpredictable human and meteorologic disturbance gradient. The rates of gradient change across the barrier island are neither gradual nor uniform. Therefore abrupt changes in the composition of the vegetation are a frequent occurrence, species often exhibiting a truncated distribution.

The boundaries of segments of the vegetation dominated by different species may be so sharp that there is little or no intergradation between them. However, this patterning should not be considered to be a case contrary to gradation but as a "segmentation of the fundamental vegetational continuum" (Whittaker 1956:36).

This distribution pattern is most apparent in the dune and swale community where individual species-type units might be considered as separate communities, as an integral segment of a zone community, or as positions on an ecosystemic gradient or ecocline (Whittaker 1970). In light of the unique complex-gradient from ocean to bay in the Sunken Forest area, the concept of an ecocline encompassing the entire barrier island is extremely useful in the interpretation of the vegetational patterns.

Dunes and swale

Due to the annual littoral erosion cycles at Fire Island in which much of the beach sand is moved offshore during the winter, the majority of the beach from the surf zone to the base of the primary dune is devoid of vegetation (Figs. 14, 15). The first vegetation encountered seaward of the primary dune were the annual species Salsola kali, Euphorbia polygonifolia (seaside spurge), and Cakile edentula (sea rocket) which were growing in the flotsam marking the high-water line of the previous winter's storms, 45 m from m.s.l. (Fig. 16). The drift lines on the beach and bay shore, in which these species are most abundant, may be extremely favorable sites for plant growth. Organic debris washed up on beaches decomposes rapidly, locally increasing the concentrations of nitrates and phosphates in the surrounding sand (Pearse et al. 1942). It is also probable that the organic matter of the drift line increases the water-holding capacity of the surface soil.

plants in drift line
Fig. 16. Annual plants (Salsola kaii and Cakile endentula) growing in drift line at base of the primary dune.

The seaward face and crest of the primary dune, where salt-spray deposition and sand movement are greatest, are dominated by Ammophila breviligulata (Fig. 17). Although Artemisia stelleriana (dusty miller) and Lathyrus japonicus (beach pea) exhibit their greatest abundance here, they do not contribute sizably to the total vegetative cover. On the lee side of the primary dune, Ammophila decreases as the abundance of woody species increases (Fig. 14, Appendix I).

Fig. 17. Beachgrass on seaward face of the primary dune.

The distribution of Ammophila breviligulata in the Sunken Forest area parallels that of A. arenaria in England; its vigor is greatly decreased in stable areas (Hope-Simpson and Jefferies 1966; Salisbury 1952). Marshal (1965) proposes as an explanation that burial by sand is a necessary stimulus for the production of new Ammophila roots, and since physiological senescence of the roots is a rapid process for this species, site stability may be a factor leading to decreased species vigor.

The growth responses of Ammophila and other plants in relation to burial by sand have already been discussed in the consideration of dune-building processes. The woody shrub species occurring on the leeward slope of the primary dune (Prunus maritima, Myrica pensylvanica (bay berry), Rhus radicans) also exhibit, to a limited extent, the capacity to respond to periodic sand burial, but to their greater susceptibility to salt-spray injury compared to Ammophila (Martin 1959).

On the leeward slope of the primary dune, where the vegetational cover is relatively undisturbed, a mixture of Ammophila breviligulata, Prunus maritima, Myrica pensylvanica, Parthenocissus quinquefolia, and Rhus radicans dominates (Fig. 18). Arctostaphylos uva-ursi (common bearberry) is a procumbent woody species that occurs in the most undisturbed locations in the dune and swale community (Fig. 19). However, much of the dune and swale community is subjected to disturbance which is reflected in the large amounts of bare sand and the abundance of species such as Hudsonia tomentosa which are associated with erosional sites (Fig. 20).

beach plum and bayberry
Fig. 18. Beach plum and bayberry grow in the wind-salt shadow on the back of the primary dune.

Fig. 19. Bearberry forming a dense spreading mat in stable portions of the swal.

beach heath and seaside goldenrod
Fig. 20. Hudsonia tomentosa and Solidago sempervirens growing on wind-eroded sites.

Major disturbance is associated with the bare sand road running in an east-west direction through the middle of the swale (Figs. 14, 15, 37). The direct effects of this 3-m-wide road, as well as the numerous foot trails in the area, can be seen in the analyses of the vegetation cover (Figs. 21, 22, Appendix I). In these sites of human disturbance the vegetation cover is reduced to zero. Furthermore, the dead plant remains are also eliminated. Consequently, the remains cannot serve the function of maintaining a cover resistant to the wind. The total effect of this disruption of the vegetation is not merely limited to the narrow avenues of human disturbance. These sites serve as the initiation of further erosional processes. Since the destruction of the vegetation and litter locally decreases the thickness of the surface boundary layer, the frequently intense winds tend to become channelized along the avenues of disturbance, and severe wind erosion may result in blowout channels. Although blowout channels are a frequent natural occurrence on barrier islands, in the Sunken Forest area they appear to have been initiated primarily by human activity.

Fig. 21. Transect A species cover.

Fig. 22. Transect A cover, space, and diversity.

The revegetation of these blowout areas appears to be a very slow process. The finer sands which characterize the dunes are transported by the wind out of the blowout areas, resulting in a pebbly surface soil. Hudsonia tomentosa and Solidago sempervirens are frequently the only species found in these sites. Neither these species, nor the others associated with blowouts, Lechea maritima (pinweed), Polygonella articulata (joint weed), Artemisia caudata (wormwood), and Panicum spp. (panic grass), form as dense stands, or have the potential for rapid vegetative spread, as does Ammophila on the dunes. Ammophila breviligulata is absent in these areas, apparently due to greater erosion than deposition of sand.

The entire center of the swale is dominated by the Hudsonia tomentosa species-type units of widely spaced individuals. On the northern side of the road there is an increase in the abundance of woody shrub and tree species: Prunus serotina, Pinus rigida, Juniperus virginiana, Rhus copallina (dwarf sumac), Vaccinium corymbosum, and Ilex opaca (Fig. 14). The vegetation on the southern face of the secondary dune is dominated by these species and marks the transition to the Sunken Forest community.

The overall vegetational pattern of the dune and swale community is one of increasing cover with increasing distance from the ocean (Fig. 22, Appendix I), the pattern being disrupted in areas of disturbance. Ammophila breviligulata, which is dominant on most of the primary dune, has an average cover of 25% in this area (Table 2). On the leeward slopes Prunus maritima and Myrica pensylvanica have an average cover of 40-55% where they occur. Arctostaphylos uva-ursi with an average cover of 69% within its species-type units, is most prevalent in the northern half of the swale. Pinus rigida achieves its greatest dominance on the margins of the Sunken Forest, but grows in a procumbent form in the swale but forming stands with an average cover of 93% (Fig. 23).

pitch pine
Fig. 23. A 50-year-old pitch pine growing as a spreading mat in the swale.

Table 2. Dune and swale community total vegetation cover.

Species-type unit Land surface area (%) covered by species-type unit
Average cover (%) of dominant species within species-type unit
Actual land surface (%) covered by dominant species

(a)x (b)= (c)
Ammophila breviligulata 19.1
Prunus maritima 6.3
Myrica pensylvanica 1.7
Hudsonia tomentosa 18.4
Arctostaphylos uva-ursi 14.1
Pinus rigida 1.4
Other species 9.0
Total Vegetation 70.0
Bare sand 30.0

Alpha-species diversity increases with increasing vegetation cover and distance from the ocean (Fig. 22). Alpha-diversity increases from less than 2 on the seaward face of the primary dune, where Ammophila attains its greatest dominance, to over 8 m portions of the swale and secondary dune. Species diversity, H(s), as calculated from species cover and the formula of Lloyd and Gherlardi (1964) shows the same patterns as alpha-diversity. H(s)-diversity varies between 0.00 and 0.25 on the seaward face of the primary dune. In the swale and on the secondary dune H(s) is generally between 1.0 and 2.0. In the areas of disturbance, particularly those adjacent to the road, both indices of diversity approach zero.

The trend of increasing cover with distance from ocean is also evident in the transect data of Martin (1959) from Island Beach State Park, New Jersey. However, this pattern is not consistent over the entire dune and swale community in the Sunken Forest area. In areas of increased disturbance, in the middle of the swale, Hudsonia tomentosa forms low density stands with an average cover of 9%. The cover and patterning of species-type units in the dune and swale community appear to be a function of the effects of a complex gradient rather than a single environmental factor such as salt spray.

The transition to the Sunken Forest community is marked by the increasing dominance of Juniperus virginiana, Pinus rigida, Prunus serotina, and Rhus copallina, which are most abundant along the margins of the forest (Fig. 24). Also increasing in abundance are Ilex opaca, Sassafras albidum, Amelanchier canadensis, and Vaccinium corymbosum, which are dominant species in the center of the forest. The physiognomy of the Sunken Forest is similar to other maritime forest communities (Wells 1939; Bourdeau and Oosting 1959; Martin 1959). The southern edge of the forest on the windward face of the secondary dune exhibits the typical espalier form found in the coastal areas (Fig. 11). Leeward of the secondary dune, the canopy rapidly assumes a relatively uniform height of 5.5-7.0 m. The tops of the trees, subjected to the shearing action of salt-laden winds, develop a smooth and closed canopy (Fig. 25). The height of the trees is ultimately determined by the height of the secondary dune which governs the wind flow patterns.

Fig. 24. The crest of the secondary dune, showing salt-spray formed canopy.

Fig. 25. The salt-pruned, flat top of the sunken Forest canopy viewed from the secondary dune crest. The Great South Bay is in the background.

Due to the relatively low canopy there is no clear stratification of the forest community into distinct overstory and understory layers. Stratification in the forest is further obscured by the abundant lianas, Rhus radicans, Smilax rotundifolia, Parthenocissus quinquefolia, Smilax glauca (sawbrier), and Vitis spp. (grape), which extend into the canopy. It is extremely difficult to give cover estimates to lianas in the canopy; however, with a mean density of 3 stems/m2 Smilax rotundifolia must be considered an important species in the forest (Table 3).

Table 3. Sunken Forest herb analyses.

Species Average %
cover ± S.E.
% Total

Rhus radicans10.2 ± 1.9 24% 88
Aralia nudicaulis8.3 ± 2.2 19 59
Gaylussacia baccata4.3 ± 1.6 10 32
Parthenocissus quinquefolia3.7 ± 0.6 9 91
Vaccinium corymbosum2.4 ± 1.0 6 18
Rosa rugosa1.5 ± 0.9 3 8
Sassafras albidum1.4 ± 0.2 3 76
Rhus copallina1.4 ± 0.8 3 9
Smilacina stellata1.3 ± 0.5 3 56
Ilex glabra1.3 ± 0.9 3 18
Maianthemum canadense1.3 ± 0.6 3 38
Pteridium adquillinum0.8 ± 0.3 2 26
Nyssa sylvatica0.7 ± 0.4 2 26
Amelanchier canadensis0.7 ± 0.2 2 47
Ilex opaca0.6 ± 0.6 1 6
Vaccinium macrocarpon0.6 ± 0.6 1 3
Vitis sp.0.6 ± 0.6 1 3
Pyrus arbutufolia0.5 ± 0.2 1 29
Trientalis borealis0.3 ± 0.1 1 18
Rhododendron viscosum0.2 ± 0.1 * 9
Phragmities communis0.2 ± 0.2 * 3
Baccharis halimifolia0.2 ± 0.2 * 6
Xanthium echinatum0.2 ± 0.1 * 3
Osmunda cinnamomea0.1 ± 0.1 * 6
Solidago sempervirens0.1 ± 0.1 * 6
Myrica pensylvanica0.1 ± 0.1 * 9
Prunus serotina0.1 ± 0.1 * 26
Teucrium canadense0.1 ± 0.1 * 3
Hypericum virginicum* * 3
Pluchea purpurascens* * 3
Gaultheria procumbens* * 3
Dryopteris thelypteris* * 3
Convolvulus sepium* * 3
Quercus velutina* * 3
Rubus sp.* * 3
   Total43.2 ± 3.3

Average density

Smilax glauca0.17 ± 0.04
Smilax rotundufolia3.03 ± 0.48

* = <0.1%

The forest is dominated by Ilex opaca, Sassafras albidum, and Amelanchier canadensis (Fig. 26). These species comprise 80% of the tree basal area (Table 4). Nyssa sylvatica occurs frequently (in 26% of the plots), but its distribution is generally limited to damp depressions which form fresh-water bogs. These four tree species are represented in the herb and shrub layer as well, but the majority of individuals are actually root sprouts. Therefore, the reproduction of the dominant tree species appears to be by vegetative rather than sexual means.

Fig. 26. American holly, black gum, sassafras, and shadbush trees in the Sunken Forest.

Peterken (1966) noted that the regeneration of Ilex aquifolium (English holly) from seed in New Forest, England, is limited to clearings and a marginal band around the edge of the forest. In the Sunken Forest there are numerous I. opaca seedlings which have germinated from seeds deposited near the bases of the female trees. However, no I. opaca seedlings older than one year were seen anywhere in the Sunken Forest during the period of this study (1967-69), indicating a high mortality rate for holly seedlings.

Table 4. Sunken Forest tree analysis (mean ± S.E.)

Species Average
(stems/100 m2)
% of total
Basal area
(cm2/100 m2
% of total
basal area

Ilex opaca5.16.2 ± 1.126873 ± 1413876
Sassasfras albidum4.430.0 ± 0.412515 ± 972391
Amelanchier canadensis4.48.1 ± 1.334435 ± 801976
Nyssa sylvantica5.30.8 ± 0.33146 ± 66626
Quercus velutina6.10.1 ± **113 ± 7959
Q. stellata6.60.1 ± 0.1*66 ± 4636
Vaccinium corymbosum2.94.0 ± 0.91650 ± 11265
Juniperus virginiana4.20.1 ± 0.1*14 ± 1413
Pinus rigida6.0**13 ± 1313
Prunus serotina3.40.3 ± 0.1111 ± 7118
Rhus copallina3.60.4 ± 0.2210 ± 6*21
Pyrus arbutifolia3.30.4 ± 0.124 ± 2*26
Rhododendron viscosum3.10.4 ± 0.213 ± 2*15
Baccharis halimifolia2.50.2 ± 0.212 ± 2*3
Rhus radicans6.00.1 ± **1 ± 1*6
Ilex glabra3.00.1 ± 0.1*1 ± 1*3
Rhus vernix3.90.1 ± 0.1*1 ± 1*3
Quercus coccinea3.8* ± ** ± *3
24.3 ± 2.0
2259 ± 202

* = *lt;0.1%

The shrub layer is dominated by Vaccinium corymbosum, Amelanchier canadensis, and Pyrus arbutifolia which constitute 70% of the basal area of the stratum (Table 5). The dominance of the shrub layer is greatest around the margins of the forest and in boggy areas. With the exception of Sambucus canadensis (common elder), Viburnum dentatum (southern arrowwood), and Rhus vernix (poison sumac), which occur in less than 5% of the plots, all the species in the shrub layer were also represented in the herb layer (Table 3) and most of them also appeared in the tree layer (Table 4).

In the herbaceous layer, Rhus radicans, Aralia nudicaulis (wild sarsaparilla), Gaylussacia baccata (black huckleberry), Parthenocissus quinquefolia, and Vaccinium corymbosum contribute over two-thirds of the total cover (Table 3, Fig. 27). The total herbaceous cover amounted to 43%, which is slightly higher than the total cover in the dune and swale community (29%). Many of the herb species occurring in the Sunken Forest community, in contrast to the dune and swale community, have wide distributions in regions removed from the barrier islands.

wild sarsaparilla
Fig. 27. Aralia nudiaulis in bloom in the Sunken Forest.

The tree canopy cover during the summer is 80-95% closed over much of the Sunken Forest, but is locally reduced around fresh-water bogs and in disturbed areas. The estimated average mid-summer forest canopy was 66%, but its cover and composition vary continually throughout the year. The phenology of the herbaceous layer seems to be correlated with seasonal changes in the forest canopy.

Different leaf-fall patterns are exhibited by Ilex opaca, Sassafras albidum, and Amelanchier canadensis (Fig 28). Sassafras leaves begin to yellow in early September, reach a peak of leaf-fall by late October, and all the leaves have fallen by early November. Amelanchier shows somewhat the same pattern but leaf-yellowing is later and leaf-fall is not completed until mid-November.

Fig. 28. Annual litter fall.

Table 5. Sunken Forest shrub analysis (mean ± S.E.)

Species Density
(stems/100 m2)
% Density Basal area
(cm2/100 m2
% B.A. Frequency

Vaccinium corymbosum3.7 ± 10.333221 ± 644465
Amelanchier canadensis9.1 ± 3.99101 ± 411741
Pyrus arbutifolia11.2 ± 4.61148 ± 18938
Rhododendron viscosum10.7 ± 7.91038 ± 3076
Baccharis halimifolia0.6 ± 0.6126 ± 2656
Ilex glabra7.4 ± 5.0721 ± 16421
Ilex opaca1.1 ± 0.6117 ± 12312
Rhus radicans11.0 ± 3.81012 ± 4241
Gaylussacia baccata7.1 ± 3.1710 ± 5218
Rhus copallina1.9 ± 1.9210 ± 1023
Sassafras albidum1.2 ± 0.517 ± 4115
Rosa rugosa5.5 ± 3.554 ± 319
Sambucus canadensis0.7 ± 0.712 ± 2*3
Nyssa sylvantica1.1 ± 0.612 ± 1*12
Viburnum dentatum0.5 ± 0.5*2 ± 2*3
Rhus vernix0.2 ± 0.7*1 ± 1*3
Pteridium aquillinum2.1 ± 2.121 ± 1*3
Prunus serotina0.1 ± 0.1***3
Rubus sp.0.1 ± 0.1***3
   Total105.7 ± 24.0
505 ± 117

* = *lt;0.1%

Ilex opaca is an evergreen species with thick opaque leaves. During the autumn and winter months, I. opaca loses some leaves during periods of high winds, but the major leaf-fall is from the middle of May to early June. This species never loses all of its leaves in a single year, and occasionally retains some leaves for as long as 3 years. The production of new I. opaca leaves in early to mid-May precedes maximum leaf-fall by about a week and leaf extention continues into June. The leafing-out of Amelanchier in late April precedes that of Ilex by 2 weeks and the leafing-out of Sassafras in mid-May follows by a week. As a result, in mid-May the canopy cover of the Sunken Forest is at a minimum.

The herbaceous layer develops rapidly from mid-May at which time Maianthemum canadense (wild lily-of-the-valley) and Smilacina stellata (false Solomon's seal) are in flower (Fig. 29). The development of the herbaceous layer appears to be essentially complete by early June when the canopy cover is nearing a maximum.

False solomons-seal
Fig. 29. Smilacina stellata in bloom in the Sunken Forest.

The phenology of flowering for the three dominant tree species also differs. Flowering precedes leaf extension in Amelanchier and reaches a peak in early May. The peak of flowering for Sassafras is in late May and early June, while Ilex flowers most abundantly in mid- to late June. The female Ilex flowers, which are pollinated by bees, develop into fruits which turn red by late September or early October. Fruit fall occurs through the autumn and into the winter, the release of fruit appearing to be greatest during periods of high wind intensity. The production of fruits by Amelanchier and Sassafras appears to be rather erratic; abundant fruit was produced in 1967 and 1968 by these species but very few fruits were produced in 1969.

In the local boggy areas within the Sunken Forest, the greatest canopy closure occurs between mid-June and late August. In early September Nyssa sylvatica leaf-fall commences and is completed by the end of that month. The vegetation of these fresh-water bogs differs greatly from the rest of the forest. Here the shrub layer is dominated by Rhododendron viscosum (swamp honeysuckle) and Vaccinium corymbosum (Fig. 30). The most common species of the herb layer in these areas are Hypericum virginicum (marsh St. John's wort), Osmunda cinamomea (cinnamon fern), O. regalis (royal fern), Nyssa sylvatica, and Dryopteris thelypteris (marsh fern).

swamp honeysuckle
Fig. 30. Swamp honeysuckle in the forest-salt marsh ecotone.

The dominant species of the bogs, except for the Osmunda spp., are also abundant in the transition zone between the Sunken Forest and the salt marshes. This ecotone, dominated by tall shrub-sized Rhododendron viscosum, Rhus radicans, Vaccinium corymbosum, Pyrus arbutifolia, and Phragmities communis (reed), is subjected to periodic flooding by salt water during exceptionally high tides. Dead Nyssa sylvatica, Juniperus virginiana, Sassafras albidum, and Quercus spp. are frequently found in the ecotone; however, very few large living trees occur there, suggesting an encroachment of the salt marsh upon the forest margin, or a slow oscillation of the ecotone (Fig. 31).

black gum
Fig. 31. Dead and dying Nyssa sylvatica in salt marsh-forest ecotone.

Salt marsh

Depending largely on topographic conditions, the width of the forest salt marsh ecotone varies from 5 to 50 m. With increasing proximity to the bay, species characteristic of brackish substrates, Pluchea purpurascens (marsh fleabane), Baccharis halimifolia (sea myrtle), Iva frutescens (marsh elder), increase in abundance. The salt marshes in the Sunken Forest area are actually pans in that they are separated from regular twice daily tidal flooding by a ridge of sand paralleling the bay shore (Fig. 32). During storms and periods of exceptionally high tide, the salt marshes are flooded with salt water. As water evaporates from the surface of the marsh, there is a concomitant increase in the surface salinity (Chapman 1964).

salt marsh
Fig. 32. Salt marsh. View is from the margin of the Sunken Forest looking toward the Great South Bay.

The vegetation of the salt marsh studied appeared to be organized into concentric zones around the standing water at its center (Table 6 and Fig. 33). Some species completely encircled the open water, Salicornia europea (chicken claws), Scirpus americanus (three-square), Rhus radicans, and Hibiscus palustris (swamp rose mallow), while other species were found only on the forest side, Rosa rugosa (rugose rose) and Pluchea purpurascens, or on the bay side Solidago sempervirens and Phragmities communis. The vegetational cover decreased gradually from the edges toward the center of the marsh where a minimum of 15% occurred.

Fig. 33. Salt marsh transect, species cover.

The vegetational changes along the salt-marsh transect were even more striking than those of the dune and swale. The boundaries between zones appear to be extremely sharp; the transition from vegetation completely dominated by Scirpus americanus to vegetation dominated by Salicornia europea takes place over a distance of 4 m (Fig. 33). However, the distributions of individual species show a tapering reminiscent of the distributions of tree species along gradients in the Great Smokey Mountains (Whittaker 1967). It is probable that in the Sunken Forest area salt marshes, the distribution of species is influenced largely by gradients in soil water and salinity (Adams 1963).

Table 6. Transect Y—salt marsh (% cover).

Species Distance from forest margin (m)
123 456 789 101112 131415 161718 192021 222324 252627 282930 313233

Hibiscus palustris
2556 4

2518 12

Scripus americanus10353540 252530505540 20105


Pulchea purpurascens
71213 22205*

Rhus radicans408


Rosa rugosa5

Baccharis halimfolia21


Salicornia europea

*1 824
15 204045353015 6045551 531618

Spartina patens


Iva frutescens



Atriplex patual


Solidago sempervirens

2830354321 14

Phragmities communis

302520121115 125080706040
Chenopodium album

5 11

Convolvulus sepium


Teucrium canadense


Cakile edentula


Open water


* = <1%

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

Last Updated: 21-Oct-2005