Abstract Question

Yüklə 0,78 Mb.
Pdf görüntüsü
ölçüsü0,78 Mb.
  1   2   3   4   5   6   7

- Structure and composition of vegetation along an elevational gradient in Puerto Rico -


Journal of Vegetation Science 17: 563-574, 2006

© IAVS; Opulus Press Uppsala.


Question: What are the composition, conservation status, and

structural and environmental characteristics of eight mature

tropical forest plant communities that occur along an elevational


Location: Northeastern Puerto Rico.

Methods: We quantified the species composition, diversity,

conservation status, and ecological attributes of eight mature

tropical forest plant communities in replicated plots located to

sample representative components of important forest types

occurring along an elevational gradient. A suite of environ-

mental and vegetation characteristics were sampled at each

plot and summarized to characterize communities and analyse

trends along the elevational gradient.

Results: The set of communities included 374 species; 92%

were native, 14% endemic, and 4% critical elements (locally

endangered) to the island. All communities, occurring within a

wide range of patch sizes and degree of conservation protec-

tion, showed a high percentage of native species (> 89% per

plot). The lowland moist forest communities, occurring within

a matrix of urbanization, agriculture, and disturbance, had the

highest degree of invasion by exotics. Community descrip-

tions were nested within a variety of hierarchies to facilitate

extrapolation of community characteristics to larger ecosys-

tem units. Basal area, above-ground biomass, canopy heights,

and mean species richness peaked at mid elevations.

Conclusions: It is significant that all of these forest communi-

ties continue to be dominated by native species while existing

in a matrix of human and natural disturbance, species inva-

sion, and forest regeneration from widespread agriculture. The

lowland moist and dry forest types represent a minority of the

protected forested areas in Puerto Rico, serve as unique ge-

netic reservoirs, and should be protected.

Keywords:  Detrended Correspondence Analysis; Exotics;

Gradient analysis; Land management; Native species; Phyto-

sociology; Puerto Rico.

Nomenclature: Liogier & Martorrell (1999).

Abbreviations:  AB = above-ground; BA = Basal area;

DCA = Detrended Correspondence Analysis;  CE = Critical

element; I = Introduced; MAP = Mean annual precipitation;

MAT = Mean annual temperature; N = Native.

Structure and composition of vegetation

along an elevational gradient in Puerto Rico

Gould, W.A.


; González, G. &  Carrero Rivera, G.

International Institute of Tropical Forestry, USDA Forest Service, Río Piedras, PR 00926-1119, USA;


Corresponding author; Fax +1 7877666302; E-mail wgould@fs.fed.us


Vegetation composition in Puerto Rico is controlled

by four key factors: climate, substrate, topography and

human and natural disturbance (Britton & Wilson 1924;

Dansereau 1966; Weaver 1991; Lugo 2005). Climatic

gradients are related to an elevational rise from sea level

up to 1340 m on the central cordillera of the island, the

northeasterly trade winds, and the rain shadow effect of

the mountains. Diverse geologic substrates include vol-

canic, limestone, and serpentine bedrock and colluvial,

alluvial, and marine quaternary deposits. Within these

broad landscape features community composition is

controlled by topographic effects on soil moisture and

development (slope position). Disturbance regimes (e.g.

land use, flooding, fire, landslides, and hurricanes) in-

duce effects on the subsequent succession of secondary

vegetation. The resulting landscape in Puerto Rico is a

mosaic of (1) primary forests that have not experienced

deforestation – according to Wadsworth (1951) and

Birdsey & Weaver (1982) <


1% of the landscape sup-

ported climax vegetation by 1950; (2) mature secondary

forests (< 10% cover of the island); (3) young secondary

forests and shrublands (> 30%); (4) agriculture, grass-

lands and pastures, (> 30%); and (5) developed areas

and urban forests, (>


15%). The mature upland forests

are primarily found in the upland forest reserves which

occupy ca. 4% of the island. These protected areas

harbour our greatest wealth of native, endemic, and

endangered plant species (Figueroa & Woodbury 1996).

There are over 2800 species of native and exotic flower-

ing plants in Puerto Rico and 9% of these are endemic to

the island (Liogier & Martorrell 1999). Understanding

the floristic composition of mature forest communities

is important for biodiversity conservation and land man-

agement planning as these serve as both reservoirs of

species and provide important ecosystem functions.

Community-level descriptions of plant associations

form the basis of hierarchical phytosociological classifi-

cation (Braun-Blanquet 1964; Mueller-Dombois &

Ellenberg 2002). This approach has been more preva-

lent in temperate than tropical vegetation studies. A few


Gould, W.A.  et al.

recent studies place tropical plant communities within the

Braun-Blanquet framework (Borhidi 1996; Bussmann

2001; Sánchez-Sánchez & Islebe 2002). While many of

the major forest formations and alliances in Puerto Rico

have been described floristically and structurally (Birdsey

& Weaver 1987; Scatena et al. 1993; Chinea 2002;

Thompson et al. 2002; Lugo 2005), the description and

classification of plant community composition within

these alliances has not been well documented.

This study has three objectives: (1) describe the

composition of eight tropical forest plant communities

within a hierarchy related to the primary controls on

vegetation, i.e. climate, substrate, topographic position,

and disturbance; (2) evaluate the conservation status

species within these communities; and (3) relate the

variation in diversity, structure, and site environmental

characteristics of these communities to gradients of cli-

mate and soil characteristics; which are correlated to



Study sites

This study was developed in Northeastern Puerto

Rico and focused on plant communities within eight

older forest types (> 60 years old) that are representative

of Lowland subtropical dry, Lowland subtropical moist,

Subtropical wet, Subtropical rain, Lower montane wet,

and Lower montane rain forest life zones in Puerto Rico

(Ewel & Whitmore 1973). The sites occur along an

elevational gradient from sea level to over 1000 m. Four

of the forest types are located in the Luquillo Experi-

mental Forest (LEF) and include Elfin woodland, Sierra

palm, Palo colorado, and Tabonuco forest alliances

(Wadsworth 1951; Weaver 1994). Four additional for-

est types include Lowland moist and dry forests, and

flooded Pterocarpus and mangrove forests. The latter

four are located on federal, state, and private managed

and unmanaged lands. All sites are on non calcareous

material derived from volcanic bedrock (montane and

lowland sites) or Quaternary deposits (coastal sites).

Field sampling

We selected 24 sampling sites throughout northeast-

ern Puerto Rico, three of each of the eight forest types.

Within these sites, 10 m 

× 10 m plots were sampled for

a total of nine replicates per forest type. Sampling was

done during 2001-2002. Replicate plots were located

away from forest edges and ecotones to represent

homogeneous stands of vegetation. The sampling

methodology was based on a priori  knowledge of

compositional differences between communities. The

plot size, while smaller than typically used to character-

ize tropical forest alliances, served well to characterize

homogeneous stands of these plant communities. Addi-

tionally, we tested whether plot sizes of 10 m 

× 10 m, 10

× 20 m, or 10 m × 30 m added a significant number of

species and found that not to be the case at our sites. The

high degree of 

β-diversity between the eight forest types

and the a priori site selection based on compositional

homogeneity within each forest type resulted in the

selection of repeatable occurrences of these communi-

ties as replicates.

Environmental and vegetation characteristics were

sampled at each plot including location, elevation, slope,

and aspect. Six soil volumes were taken from 0-10 cm

depths at each plot and combined to form a composite

sample, which was ground with a Foss Tecator Cyclotec

(model 1093) mill through a 1 mm stainless steel sieve.

Total N, C and S were determined using the dry com-

bustion method by means of a LECO CNS-200 analyzer

(St. Joseph, MI, USA). Elemental P, Ca, K, Al, Fe, Mn,

and Mg were determined from digestions using a

Beckman plasma emission spectrometer (Spectra V)

(Luh Huang & Schulte 1985). Mean annual precipita-

tion was sampled bimonthly from July 2001 to July

2004 and mean annual air temperature was determined

from shaded stations placed at 1 m height. All plant

species were identified following Liogier and Martorrell

(1999). Species cover was estimated using the Braun-

Blanquet cover scale (Westhoff & van der Maarel 1978).

Percent cover and height estimates were made for each

vegetation layer (canopy trees, understorey trees, shrub

layer, and herbaceous layer). Percent cover estimates

were made for non vegetation layers including bare soil,

root, rock, litter, and water. Tree diameters were meas-

ured at breast height (DBH) for all trees over 2 cm.

Data analyses

All species were classified as either native or non

native (Little & Wadsworth 1964). Native species were

classified as endemic and/or critical elements using the

classification developed by the Puerto Rico Department

of Natural and Environmental Resources for conserva-

tion purposes (Quevedo 1999). Species were identified

by growth form as trees, shrubs, lianas, forbs, graminoids,

or epiphytes. For trees, we determined the number of

stems per species per plot and forest type and the basal

area per species per plot. We calculated biomass per

species per plot using DBH and regression equations

from Brown (1997) for lowland subtropical dry for-

ests and Scatena et al. (1993) for all other forests.

Tree species were ranked by importance values (IV)

calculated per forest type using relative frequency (RF)

- Structure and composition of vegetation along an elevational gradient in Puerto Rico -


(occurrence within nine plots per forest type), relative

dominance (RDOM) (based on basal area from D


measurements), and relative density (RDEN) (based on

number of stems) in the following relationship

IV = (RF 

× RDOM × RDEN) / 3)


(Kent & Coker 1996). We determined the percent

constancy (or relative frequency) of all herbaceous

and woody species per community type as the percent

of plots in which it occurred in that community type.

The percent fidelity was calculated as the percent of

occurrences within a community relative to all occur-


Two tree species, the characteristic (high constancy

and importance value (IV) and diagnostic (high fidelity

and constancy) species within the data set, were used

to name the community types (except in the mono-

specific Pterocarpus forest). Each community is nested

in a series of hierarchical crosswalk tables to place the

community within a phytosociological class and order

(Braun-Blanquet 1964) and within a vegetation alli-

ance or an equivalent designation and formation (Anon.

1998; Ahern et al. 1999; Areces-Mallea et al. 1999;

Dansereau 1966). Communities were hierarchically

nested within life zone (Ewel & Whitmore 1973),

vegetation zone (Dansereau 1966), forest region

(Wadsworth 1964), and mapped units (Helmer et al.

2002). These include descriptions consistent with glo-

bal forest cover classes used by the Food and Agricul-

ture Organization (Anon. 1998) and the Global Obser-

vations of Forest Cover (Ahern et al. 1999).

Ordinations were performed on the 54 non wetland

samples using PCORD 4.0 (McCune & Medford 1999),

including all species and using Braun-Blanquet cover

estimates from each plot as abundance measures

(Westhoff & van der Maarel 1978). Univariate analy-

sis of means and Student-Newman-Keuls (S-N-K) (


= 0.05) were used to determine whether the commu-

nity types differed significantly in the numbers of

native, introduced, endemic, and endangered species,

soil attributes, and structural characteristics (Anon.





We found 374 plant species from 98 families within

the plots. The majority were woody plants. The woody

component was primarily composed tree species, fol-

lowed by woody vines (lianas) and shrub species. Nearly

half of all tree species were found with diameters greater

than 10 cm. Over 75% of all tree species were found

with diameters greater than 2 cm and 85% of all tree

species were found with diameters less than 2 cm (sap-

lings). Over half of the herbaceous component was

made up of forbs, followed by lesser numbers of epi-

phytes, and graminoids (Table 1).

Of the species sampled 93% were native. Endemics

make up 14% of the species sampled and 4% are critical

elements (Quevedo 1999). Exotic species include 13

introduced tree species with diameters > 2 cm and most

(with the exception of Hura crepitans in the Lowland

moist forest) were minor elements within the communi-

ties (App. 1).

Community classification

Eight plant community types will now be described

and as far as possible phytosociologically identified on

the basis of the vegetation description of Cuba by Borhidi


Table 1.  Number and percent of all species and of subgroups

by growth form, stem size, and conservation status for all sites.




all species

by subgroups

Plant species



Woody species



Tree species



65 of woody spp.


≥ 10 cm



44 of tree spp.

  Trees  > 2 cm DBH



76 of tree spp.

  Trees < 2 cm DBH (saplings) 147


85 of tree spp.

Shrub species



8 of woody spp.

Woody vines (lianas)



27 of woody spp.

Herbaceous species






52 of herbaceous spp.




11 of herbaceous spp.




38 of herbaceous spp.

Native species



Non endemic



85 of native spp.




15 of native spp.

    Critically endangered



4 of native spp.

Exotic species




Gould, W.A.  et al.

1. The Tabebuia rigida-Eugenia borinquensis  com-

munity (App. 1a, Tables 2-5)

This is a closed broad-leaved evergreen forest

(Anon. 1998; Ahern et al. 1999) found within the

Elfin woodland forest type or alliance (Weaver 1994)

and Montane subtropical cloud forest formation (Ta-

ble 2). It is within the class Cyrillo-Weinmannietae

pinnatae, order Cyrillo-Weinmannietalia pinnatae

(Borhidi 1996) of tropical cloud forest vegetation

(Table 2). The community is found within both the

Lower montane wet and rain Holdridge life zones

(Ewel & Whitmore 1973, Table 3) on azonal sites

including exposed slopes and ridges (Lyford 1969;

Weaver 1990, 1991,1994; Weaver et al. 1986). It is a

gnarled, epiphyte-laden, and dense forest type char-

acterized by closely spaced, stunted trees. The soils

are commonly saturated with moisture as the forest is

frequently enveloped in clouds. Seven of the ten tree

species with DBH > 2 cm are endemic. Rare endemics

include  Ardisia luquillensis  and  Ilex sintenissi. The

endemic shrub Miconia foveolata was found in 44%

of the plots. The lianas Marcgravia sintenissi  and

Mikania cordifolia are characteristic of the commu-

nity, as are the endemic forbs Cyathea borinquena

and Pilea yunquensis. The grass Isachne angustifolia

occurred at most sites as did endemic epiphytes

Selaginella krugii, Hohenbergia attenuata and native

epiphytic fern Blechnum fragile.

2. The Prestoea montana-Cecropia schreberiana com-

munity (App. 1b, Tables 2-5)

This is a closed broad-leaved evergreen forest

(Anon. 1998; Ahern et al. 1999) found within the

Sierra palm forest type or alliance (Weaver 1991,

1994) and Montane subtropical rain forest formation

(Table 2). It is provisionally placed within the class

Ocoteo-Cyrilletea racemiflorae and order Ocotea-

Magnolietalia  (Borhidi 1996) of montane tropical

rainforests (Table 2). The Sierra palm forest type is

found in the Lower montane wet and subtropical rain

Holdridge life zones (Ewel & Whitmore 1973, Table

3). The communities are found on azonal (although

common) steeply sloping sites and dominated by

Prestoea montana (Sierra palm ) and an abundance of

epiphytes. The abundance of Cecropia schreberiana

in this community is indicative of past hurricane

disturbance. Five endemic tree species with DBH > 2

cm occur at the sites, the most abundant being

Henrietta squamulosa, Eugenia borinquensis  and

Cyathea arborea. Two endemic lianas characterize

the community including Macgravia sintenissi and

Ipomoea krugii. Eleven forbs were consistently found

within the community including two endemics

Cyathea borinquena and Begonia decandra. The grass

species Ichnanthus pallens and epiphytes Guzmania

berteriana and Selaginella krugii were also charac-


3. The Cyrilla racemiflora-Micropholis garciniifolia

community (App. 1c, Tables 2-5)

This is a closed broad-leaved evergreen forest

(Anon. 1998; Ahern et al. 1999) found within the

Palo colorado forest type or alliance (Weaver 1986,

1991, 1994) and Montane subtropical rain forest for-

mation (Table 2). It is within the class Ocoteo-

Cyrilletea racemiflorae and order Ocotea-Magno-

lietalia  (Borhidi 1996) of montane tropical rainfor-

ests (Table 2). It is characterized by the presence of

Cyrilla racemiflora (Palo colorado) and occurs in the

Lower montane wet Holdridge life zone and is con-

sidered the zonal vegetation of this life zone (Ewel &

Whitmore 1973, Table 3). The sites included 14 en-

demic tree species and 6 rare native species with

DBH > 2 cm. Saplings of Prestoea montana,

Psychotria maleolens, Myrcia fallax and  Hirtella

rugosa were present in most sites. The endemic liana

Marcgravia sintenissi was present in most sites, as

were the ferns Thelypteris deltoides, Cyathea

borinquena,  and  Nephrolepis rivularis,  the grass

Ichnanthus pallens, and epiphytes Anthurium

dominicense and Polypodium dissimile.

4. The Dacryodes excelsa-Manilkara bidentata com-

munity (App. 1d, Tables 2-5)

This is a closed broad-leaved forest (Anon. 1998;

Ahern et al. 1999) found within the Tabonuco forest

type or alliance (Weaver 1989, 1994; Basnet 1992)

and Submontane subtropical rain forest formation

(Table 2). It is provisionally placed within the class

Swietenio-Calophylletea calabae and order Dipholi-

Calyphylletalia calabae (Borhidi 1996) of submontane

tropical rain forests (Table 2). It is characterized by

the presence of Dacryodes excelsa (Tabonuco ) and

occurs in the Subtropical wet Holdridge life zone

(Ewel & Whitmore 1973, Table 3). Tetragastris

balsamifera  is also common and was found in over

40% of the plots in the overstorey and as saplings in

78% of the plots. The community includes five en-

demic tree species, three considered rare (App. 1d).

Twelve species occurred as saplings in > 50% of the

plots, two of these restricted to the tabonuco plots

include Myrcia leptocladia and Eugenia stahlii. Four

lianas occurred in most plots, including Rourea

surinamensis in all but one tabonuco plot. The ferns

Cyathea borinquena  and  Thelypteris deltoides  oc-

curred in all plots and the grass Ichnanthus pallens

occurred in nearly all plots.

- Structure and composition of vegetation along an elevational gradient in Puerto Rico -


5. The Manilkara bidentata-Ocotea leucoxylon com-

munity (App. 1e, Tables 2-5)

This is a closed broad-leaved forest (Anon. 1998;

Ahern et al. 1999) found within the Lowland rain

forest alliance (Dansereau 1966) and Lowland sub-

tropical seasonal evergreen forest formation (Table

2). It is placed within the class Swietenio-Calophyl-

letea calabae and order Dipholi-Calophylletalia

calabae (Borhidi 1996) of submontane tropical rain-

forests (Table 2). It is likely similar to formerly ex-

tensive moist lowland forests converted to agricul-

ture in the preceding century. It occurs in the Low-

land subtropical moist Holdridge life zone (Ewel and

Whitmore 1973, Table 3). Manilkara bidentata  is

dominant in the overstorey of most plots, Faramea

occidentalis is the most abundant understorey species

and Inga laurina was common in all plots. The low-

land moist plots were more variable than the montane

plots and Ocotea leucoxylon and Hymenaea courbaril

shared dominance with or replaced Manilkara

bidentata in some plots. One rare endemic tree species,

Thespesa grandiflora, occurred as well as one rare

native,  Byrsonima spicata. Four lianas occurred at

most sites including Macfadyena unguis-cati, Paul-

linia pinnata, Hippocratea volibis and Rourea surina-

mensis. The fern Adiantum latifolium occurred in all

plots and the grasses Ichnanthus pallens and Lasiacis

divaricata occurred in most plots.

6. The Bucida buceras-Guapira fragrans community

(App. 1f, Tables 2-5)

This is a closed mixed-evergreen deciduous forest

(Anon.  1998; Ahern et al. 1999) found within the

Bucida buceras (Ucar) forest alliance (Dansereau

1966) and Lowland subtropical semideciduous wood-

land formation (Table 2). It is placed within the class

Guazumo-Ceibetea pentandrae and order Oxandro-

Burseretalia (Borhidi 1996) of semideciduous meso-

phytic forests (Table 2). It occurs on non calcareous

substrates in the Lowland subtropical dry Holdridge

life zone (Ewel & Whitmore 1973, Table 3). Bursera

simaruba  and  Bourreria succulenta  also occurred

and were abundant in most plots. The rare native

Maytenus cymosa occurred in one plot and the sap-

lings  Capparis cynophallophora, Erythroxylum

brevipes, Neea buxifolia, and Eugenia biflora  oc-

curred in most plots. The lianasMacfadyena unguis-

cati, Tragia volubilis, Serjania polyphylla, and

Chioccocca alba  occurred in most plots as did the

grass Ichnanthus pallens.

7. The Pterocarpus officinalis-Acrostichum aureum

community (App. 1g, Tables 2-5)

This is a closed flooded forest (Anon. 1998, Ahern

et al. 1999) found within the Pterocarpus officinalis

forest alliance (Dansereau 1966) and Subtropical sea-

sonally flooded rain forest formation (Table 2). It is

placed within the class Chrysobalano-Annonetea

glabrae and order Chrysobalano-Annonetalia glabrae

(Borhidi 1996) of freshwater swamp forests (Table

2). It occurs in relict patches within the Lowland

subtropical moist Holdridge life zone (Ewel and Whit-

more 1973, Table 3) and was likely much more exten-

sive on the coastal plain and riparian corridors prior

to deforestation for agriculture (Alvárez 1982, Eusse

& Aide 1999). It is found on undisturbed freshwater

seasonally flooded non calcareous alluvial substrates.

It is typically a monoculture of Pterocarpus trees in

the overstory with the fern Acrostichum aureum  in

the understory. The lianas Paullinia pinnata,

Hippocratea volubilis,  and  Ipomoea triloba  were

found on most plots.

8. The Avicennia germinans-Laguncularia racemosa

community (App. 1h, Tables 2-5)

This is a closed flooded forest (Anon. 1998; Ahern

et al. 1999) found in the Mixed mangrove alliance

(Dansereau 1966) and the Tidally flooded evergreen

sclerophyllous (mangrove) forest formation (Table

2). It is placed within the class Rhizophoro-

Avicennietea germinantris and order Avicennietalia

(Borhidi 1996) of mangrove forests (Table 2). Man-

grove forest types occur within the Lowland sub-

tropical dry and moist Holdridge life zones (Ewel &

Whitmore 1973, Table 3) in Puerto Rico and are

typically found as Rhizophora mangle  dominated

stands on coastal and estuarine fringes and in basins

as pure or mixed stands of Avicennia germinans and

Laguncularia racemosa. Nearly all plots of this com-

munity had both Avicennia  germinans  and  Lagun-

cularia  racemosa,  with  Avicennia  germinans  typi-

cally being the dominant. The shrub Suriana maritima

occurred in most plots.


Gould, W.A.  et al.

Yüklə 0,78 Mb.

Dostları ilə paylaş:
  1   2   3   4   5   6   7

Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©www.azkurs.org 2020
rəhbərliyinə müraciət

    Ana səhifə