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

NPS Scientific Monograph No. 7
NPS Logo


The ecocline from the beach to the forest is characterized by increases of plant cover, organic matter on the soil surface, height of the dominant vegetation, and amelioration of extremes in the environment. This spatial sequence strongly resembles the temporal changes associated with vegetational succession. The resemblance has led some (Gooding 1947; Kurz 1939) to consider the sequences as successional stages or to interpret physiognomy as being indicative of the age of the vegetation (Davis 1957). However, there was no evidence of active succession between the community zones in the Sunken Forest area.

Changes in the ocean and bay shores will obviously cause readjustments in the vegetation of the adjacent areas, but at present this process does not appear to be affecting the forest proper. In the Great South Bay, Juniperus virginiana stumps are visible 10 m offshore during low tides, giving evidence of long-term bay shore erosion (Fig. 34). Likewise, the seaward face of the primary dune is subjected to periodic wave erosion, a process that appears to be accelerating.

Fig. 34. Stumps in the Great South Bay are exposed at low tide. Winter storm deposition of dead eelgrasss can be seen on the boardwalk in the background.

However, at Fire Island there is no evidence suggesting recent large-scale changes in the position of the Sunken Forest. No stumps indicative of previous forest growth have been found on the beach as in North Carolina (Brown 1959). There is no evidence that the secondary dune system is encroaching upon the forest. Any major changes in the dimensions of the forest have probably been in the ecotone with the salt marshes, where standing dead trees are common; but it is difficult to assess the magnitude or permanence of changes in this area.

As on other barrier islands, the sequence of plants from ocean to forest or from bay to forest does not necessarily indicate forest succession (Martin 1959; Oosting 1954). The community zones appear to be spatially stable relative to each other, but changes in the complex-gradient lead to readjustments indicative of succession within the zones. Any point in the complex-gradient can exhibit succession, but terminal development is limited to a maximum for that position on the gradient.

Due to the recreational development of the west end of Fire Island, it is difficult to find sites for the investigation of primary succession. Some indication of primary succession on bare sand was gained through the examination of an island constructed in the Great South Bay by the National Park Service from dredge material (Fig. 35). The construction of this island 50 m in diameter and 2 m relief took place in the autumn of 1966, and no vegetation was observed during the summer of 1967. A reexamination in the autumn of 1969 showed a large number of annual and perennial herbaceous species: Salsola kali, Chenopodium album (lamb's quarters), Xanthium echinatum (sea burdock), Phragmities communis, Scirpus americanus, Arenaria peploidies (sea beach sandwort), Cakile endentula, Ammophila breviligulata, Amaranthus albus (pigweed), Lycopersicum esculentum (tomato), and Cucumis sativus (cucumber). These plants had been disseminated by both water and birds, except for the last two species which were undoubtedly of an anthrochorous nature.

Dredge Island
Fig. 35. Dredge Island off Watch Hill, Fire Island.

A further resurvey of the dredge island during the summer of 1973 indicated an invasion by Spartina patens (salt marsh hay), Digitaria (crabgrass), Solidago sempervirens, Teucrium canadense (wood sage), Convolvulus sepium (hedge bindweed), Baccharis halimifolia, Cyperus strigosus, and Spartina alterniflora (salt marsh grass). The following species were no longer found: Arenaria peploides, Ammophila breviligulata, Lycopersicum and Cucumis. The demise of Ammophila may have been due to the lack of sediment sources for deposition on the island.

The vegetational development of this sand island is not directly comparable to succession on the barrier island proper, since the former is located in the Great South Bay and therefore is not subjected to the environmental conditions of the Atlantic shore.

Although at present the community zones in the Sunken Forest area appear to be in equilibrium with factors of the environmental complex gradient, the historical development of the maritime forest involved changes in dominant species. If one assumes, as Martin (1959) does, that the secondary dune was formed by processes identical to those that build the primary dunes, then the earliest stages in succession in the development of the maritime forest must have resembled the dune and swale community at present. Pollen analyses also suggest a pioneer dune vegetation preceded the present vegetation in the Sunken Forest (Sirkin 1972).

In areas of the dune and swale where sand is being deposited Ammophila breviligulata is a pioneer plant (Fig. 36). Ammophila is displaced by a variety of woody species including Prunus maritima, Myrica pensylvanica, and Arctostaphylos uva-ursi, except on the seaward face of the primary dune where Ammophila must be considered a "climax" species.

Fig. 36. Successional trends in the Sunken Forest area.

In blowouts where erosion is an active process, Hudsonia tomentosa appears to be the main pioneer species. Arctostaphylos uva-ursi, growing on the margins of the blowouts, spreads vegetatively into the Hudsonia stands and eventually displaces the species (Fig. 37). A variety of species such as Rhus copallina, Smilax glauca, and Aralia nudicaulis invade the dense Arctostaphylos uva-ursi mats in the subsequent successional stage.

Fig. 37. Arctostaphylos (left) invading a blow-out occupied by Hudsonia (right). Sand road is in the background.

Pinus rigida and Juniperus virginiana may become established directly on relatively stable bare sand sites in the dune and swale community. Thus there appear to be three different pathways of succession depending upon stability conditions soil surface (Fig. 36).

The climax vegetation in the dune and swale north of the community primary dune crest is probably a mixture of woody shrub species exhibiting a low growth form: Prunus maritima, Myrica pensylvanica, Vaccinium corymbosum, Pinus rigida, and Juniperus virginiana. Studies by Waterman (1919) of the root development of dune species suggest that early successional species, particularly in areas of sand accretion, have root systems that exhibit negative responses to organic matter in the soil, while woody species of later successional stages have root systems that exhibit a positive response to soil organic matter. This relationship between succession and physiological responses may not be as clear-cut as Waterman indicated, for Salisbury (1952) reported finding Ammophila arenaria roots concentrated in the sand layers richer in organic matter in dunes at Norfolk, England.

The terms succession and climax are not as useful in describing directional changes of communities on barrier islands as they are in inland sites. Erratic wind and water disturbances have undoubtedly played a prominent role in shaping the vegetation of barrier islands long before human disturbance was a major factor. In such an environment of continual disturbance, the labels of "pioneer" and "climax" species are somewhat meaningless.

In the undisturbed and protected northwest section of the dune and swale community, a 35 X 60-m stand consisting of Quercus stellata (post oak), Ilex opaca, Amelanchier canadensis, and Vaccinium corymbosum has become established (Fig. 14). The presence of these forest species in the swale is dependent upon topographic features which locally alter the flow of salt-laden winds. As the plants outgrow their protection, vertical height growth is arrested by the deposition of salt spray, their resultant growth form being that of a shrub.

The development of the present maritime forest in the Sunken Forest area seems to have been primarily a function of the unbroken secondary dune system. The tree species which presently dominate the crest of the secondary dune and margins of the forest (Prunus serotina, Juniperus virginiana, Pinus rigida, and Quercus spp.) were probably dominant in the seral stage immediately preceding the present forest (Fig. 36).

At present Juniperus virginiana, Pinus rigida, and Quercus spp. populations all appear to be declining in the Sunken Forest community. Juniperus virginiana and Pinus rigida account for 44% of the dead-tree density and 63% of the standing dead-tree basal area (Table 7). More importantly, these species have a much higher ratio of dead:living individuals and basal area than the species which are currently dominant in the forest.

Table 7. Sunken Forest dead tree analysis.

Species Density (stems/100m2)
Basal area (cm2/100 m2)
LivingDead Living:Dead LivingStanding dead Living:Dead

Pinus rigida 0.03 0.29 0.2 13.3 64.6 0.2
Juniperus virginiana 0.06 0.32 0.2 13.5 6.2 2.2
Rhus vernix 0.06 0.03 2.0 0.6 0.2 3.0
Rhus radicans 0.06 0.03 2.0 1.4 0.4 3.5
Quercus stellata 0.09 0.06 1.5 66.0 7.0 9.4
Rhus copallina 0.38 0.03 12.7 9.9 0.6 16.5
Sassafras albidum 2.97 0.24 12.4 514.7 18.0 28.6
Amelanchier canadensis 8.12 0.26 31.2 435.2 8.4 51.8
Nyssa sylvatica 0.79 0.03 26.3 146.1 2.6 56.2
Ilex opaca 6.24 0.09 69.3 873.5 4.1 213.0
Total (All species) 24.27 1.38 17.6 2258.7 112.1 20.1

Table 8. Sunken Forest tree density distributions: Density (D.) stems/100 m2.

DBH Class 1
3.0-9.9 cm
DBH Class 2
10.0-19.9 cm
DBH Class 3
20.0-29.9 cm
DBH Class 4
30.0-39.9 cm
DBH Class 5
>40 cm
D.% D.% D.% D.% D.%

Ilex opaca2.4153.2480.5450.133

Sassafras albidum1.491.1170.3270.133

Amelanchier canadensis6.4391.6240.19

Nyssa sylvatica0.210.580.19

Quercus velutina


Quercus stellata


Vaccinium corymbosum4.025

Pinus rigida


Juniperus virginiana


Prunus serotina0.32**

Rhus copallina0.42**

Pyrus arbutifolia0.42

Rhododendron viscosum0.42

Baccharis halimfolia0.21

Rhus radicans0.11

Ilex glabra0.11

Rhus vernix0.11

Quercus coccinea*1


* = <0.1%

Table 9. Sunken Forest tree basal area distributions: Basal Area (B.A.) cm2/100 m2.

Species DBH Class 1
3.0-9.9 cm
DBH Class 2
10.0-19.9 cm
DBH Class 3
20.0-29.9 cm
DBH Class 4
30.0-39.9 cm
DBH Class 5
>40 cm
B.A.% B.A.% B.A.% B.A.% B.A.%

Ilex opaca99.524503.448219.34942.317

Sassafras albidum39.410189.818136.631149.461

Amelanchier canadensis178.544216.92140.310

Nyssa sylvatica13.73109.21023.55

Quercus velutina


Quercus stellata


Vaccinium corymbosum49.512

Pinus rigida


Juniperus virginiana


Prunus serotina7.024.3*

Rhus copallina6.023.8*

Pyrus arbutifolia4.51

Rhododendron viscosum3.21

Baccharis halimfolia2.31

Rhus radicans1.4*

Ilex glabra0.7*

Rhus vernix0.7*

Quercus coccinea0.4*


* = <0.1%

The distributions of basal area and density by strata and diameter size classes (Tables 3, 4, 5, 8, 9) indicate differences in the population structure of the forest tree species. Three groups of species can be abstracted from these distributions. The first group consists of species with individuals in the herb layer, the shrub layer, and at least three diameter size classes in the tree layer: Ilex opaca, Sassafras albidum, Amelanchier canadensis, and Nyssa sylvatica. These species can be considered the climax species under present conditions in the Sunken Forest.

The second group includes species which are concentrated in the lower strata and the smallest size class of the tree strata: Vaccinium corymbosum, Rhus copallina, Pyrus arbutifolia, Rhododendron viscosum, Baccharis halimifolia, Rhus radicans, and Ilex glabra (inkberry). These species are actually large shrubs which would not be found in larger size classes even under optimum conditions.

The third group of tree species are those absent from the tower strata and smallest tree layer diameter size classes: Pinus rigida, Juniperus virginiana, Quercus stellata, and Q. velutina (black oak). Generally, these species do not reproduce successfully in deep shade (Fowells 1965), a condition existing in much of the Sunken Forest. If the present environmental conditions were to continue unaltered, these species would eventually be eliminated from all but the margins of the forest. However, these species may actually be elements of the "climax" forest perpetuated by shifting patterns of disturbance as in the case of Acer saccharum (sugar maple) in Wisconsin maple-basswood forests (Bray 1956) and Betula allegheniensis (yellow birch) (Forcier 1973).

The occurrence of Ilex, Sassafras, and Amelanchier individuals in a wide variety of size classes, coupled with the relative absence of dead individuals of these species, suggests the Sunken Forest is approaching climax, but perhaps for the first time, the ages of the dominant trees being 100-170 years. The notion that the present Sunken Forest is a first growth Ilex-Sassafras-Amelanchier forest is supported by a radio-carbon dating of a peat sample from the bottom of a fresh-water bog. The age of the sample, taken at a depth of 3 feet, was 250 ± 80 years before the present (Sirkin 1972), suggesting that the Sunken Forest area has been stable for 200-300 years.

The diameter size classes used in Tables 8 and 9 are not truly equivalent to age classes for the trees. A series of increment cores was taken from representative dominant trees in the forest (Table 10). The ages of the dominant Ilex are generally 100-150 years, while dominant Sassafras and Amelanchier are somewhat younger. The largest Ilex opaca cored (402 mm dbh) was the oldest tree sampled in the forest (164 years). However, the dbh for the other individuals cored was a relatively poor index of their absolute age.

Exact changes that have occurred in species composition in recent years are difficult to ascertain. It is probable that Amelanchier canadensis has increased in dominance in the Sunken Forest since 1933 when Murphy mentioned it only as a shrub. Otherwise, the species composition of the dominant Sunken Forest vegetation is not markedly different from the list of component trees in their approximate order of abundance given by Murphy: Ilex opaca, Nyssa sylvatica, Sassafras albidum, Prunus serotina, Quercus spp., Pinus rigida, Juniperus virginiana, and Acer rubrum (red maple).

Table 10. Sunken Forest tree ages.

Species Diameter at
breast height (mm)
Age at
breast height

Ilex opaca 162 45
I. opaca 117 50
I. opaca 168 77
I. opaca 173 102
I. opaca 382 108
I. opaca 198 111
I. opaca 152 118
I. opaca 273 140
I. opaca 328 143
I. opaca 81 143
I. opaca 161 145
I. opaca 402 164
Sassafras albidum 173 79
S. albidum 224 88
S. albidum 174 88
S. albidum 168 91
Amelanchier canadensis 168 54
A. canadensis 143 57
A. canadensis 147 70
A. canadensis 155 72
Quercus velutina 165 28
Q. velutina 418 105
Juniperus virginiana 240 139
Pinus rigida 374 88
Nyssa sylvatica 184 79
Quercus stellata 410 108

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

Last Updated: 21-Oct-2005