Forest Ecology and Management 199 (2004) 379–393
Short-term response of secondary forests to hurricane
disturbance in Puerto Rico, USA
John B. Pascarellaa,*, T. Mitchell Aideb, Jess K. Zimmermanc
a
Department of Biology, Valdosta State University, 1500 N. Patterson Street, Valdosta, GA 31698, USA
Department of Biology, University of Puerto Rico-Rio Piedras, PO Box 23360, San Juan, PR 00931-3360, USA
c
Institute for Tropical Ecosystem Studies, University of Puerto Rico-Rio Piedras, PO Box 23341, San Juan, PR 00931-3341, USA
b
Received 5 February 2004; received in revised form 24 May 2004; accepted 24 May 2004
Abstract
We examined the short-term (4–5 years) dynamics of stand structure and species composition of two chronosequences of
secondary forest stands derived from abandoned cattle pastures (15–81 years since abandonment) in the Luquillo and Carite
Mountains of eastern Puerto Rico. In 1998, Hurricane Georges struck Puerto Rico, affecting both chronosequences.
Stem densities decreased in all sites with a significant effect of age-class. Intermediate-aged sites had the greatest decrease in
density and the largest change in the distribution of size classes. Although there was a significant decrease in the small size
classes (<10 cm dbh), due to physical damage from falling debris, there was a large increase in the 10–20 cm dbh classes, due to
growth of surviving stems. Basal area decreased in 11 of the 15 sites and showed only a slight gain in the other sites. Species
richness decreased in all but one site and there was a greater proportional loss of shrub species than tree species. Multivariate
ordination successfully placed the resampled sites within their original ordinations with acceptable stress, indicating there was
no significant change in species composition or deviation from successional pathways during the period. Compared with effects
of land-use, hurricanes create minor variation from expected successional pathways. However, hurricanes contribute to regionaland landscape-scale variation in both stand structure and species composition.
# 2004 Elsevier B.V. All rights reserved.
Keywords: Chronosequence; Forest dynamics; Hurricane; Land-use; Secondary succession; Tropical forests
1. Introduction
Chronosequences are the use of multiple sites of
varying ages to examine successional trajectories of
populations, communities, and ecosystems (Gerrish
and Mueller-Dombois, 1999; Verchot et al., 2000;
Davis et al., 2003; Erickson and Hamrick, 2003;
Lehmkuhl et al., 2003). Implicit in the chronose*
Corresponding author. Tel.: þ1-229-333-5766;
fax: þ1-229-245-6585.
E-mail address: [email protected] (J.B. Pascarella).
quence approach is the substitution of space for time,
in that the alternative approach of tracking sites
through time is too lengthy for examining long-term
successional patterns. Critics of the chronosequence
approach have focused on the lack of incorporation of
environmental variation and historical factors, producing an overly deterministic model that solely focuses
on the age of the site (Pickett, 1989). Indeed, reanalysis of chronosequence studies have produced dramatic differences of opinion regarding mechanisms
and outcomes of succession (see Fastie, 1995). Nonetheless, well-designed chronosequence studies with
0378-1127/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.foreco.2004.05.041
380
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
replicated site ages, quantitative and qualitative measurements of environmental variables, and appropriate
multivariate analyses are an important tool for examining long-term successional patterns. Although much
less common but highly valuable, the longitudinal
tracking of sites within the chronosequence can be
used to validate the predictive capability of the original chronosequence study. This tracking can examine
the observed short-term changes relative to the predicted changes and can examine the ecological causes
of deviations from the predicted changes, including
the effects of disturbance (Collins and Adams, 1983;
Twigg et al., 1989; Baker et al., 1996; Debussche et al.,
1996; Foster and Tilman, 2000).
Chronosequences have been particularly useful in
studying succession in tropical forest ecosystems
(Saldariagga et al., 1988; Aplet and Vitousek, 1994;
Grau et al., 1997; Dewalt et al., 2000; Rivera et al.,
2000; Kennard, 2002). In Puerto Rico, economic
changes during the 20th century from an agricultural-based economy to an industrialized economy
have resulted in an island-wide recovery of secondary
forest (Dietz, 1986; Birdsey and Weaver, 1987; Rudel
et al., 2000). Chronosequence studies have provided
extensive information on successional patterns of
tropical secondary forests in a variety of life-zones,
soil types, and land-use categories (Zimmerman et al.,
1995; Aide et al., 1995, 1996, 2000; Lugo et al., 1996;
Rivera and Aide, 1998; Pascarella et al., 2000; Marcano-Vega et al., 2002). These studies have emphasized the importance of age since abandonment on
forest structure, and age and elevation on species
composition (Aide et al., 1996; Pascarella et al.,
2000), but there has been little consideration to the
impact of hurricanes on secondary forest succession.
Hurricanes are important natural disturbances in the
Caribbean. Hurricanes strike Puerto Rico on average
every 22 years and have significant effects on ecosystem processes, vegetation, and animals (Brokaw and
Walker, 1991; Lodge and McDowell, 1991; Tanner
et al., 1991; Waide, 1991). For forest trees, hurricanes
cause moderate to severe defoliation and can directly
damage and kill trees through uprooting, breakage and
loss of minor and major branches, and stem breakage
(Walker, 1991; Zimmerman et al., 1994). Furthermore, elevated inputs of nutrients, particularly nitrogen and phosphorus (Lodge et al., 1991), an increase
in understory light levels (Fernandez and Fetcher,
1991), and a lower canopy stature (Brokaw and Grear,
1991) can also affect vegetation dynamics. In mature
forests, these direct and indirect effects result in
changes in species composition (Crow, 1980) that
may significantly influence the successional process.
The major objective of this study was to determine
how the predicted trajectory of forest succession,
based on a chronosequence, is affected by a major
disturbance (e.g., a hurricane). If hurricanes have a
negligible impact on succession, we hypothesized
that the sites would behave as predicted by the
chronosequence model (Aide et al., 2000; Pascarella
et al., 2000) with the following changes: (1) an
increase in stem density in young sites, a peak in
intermediate aged sites, and a decrease in older sites,
(2) a rapid but then declining positive rise in basal
area, (3) a rapid increase in species richness that
peaks at intermediate-aged sites and then slowly
declines, and (4) a minimal change in the species
composition of a site, with the largest change occurring in the young sites. In contrast, if hurricanes have
a large impact on succession, we hypothesized that
there would be: (1) a decrease in stem density and
basal area from the death of killed trees, particularly
in sites with large trees and in sites with an abundance
of weak-wooded pioneer trees, (2) a minimal change
in species richness with the loss of established species compensated by the recruitment of pioneer species and suppressed understory species, and (3) a
substantial change in species composition within a
site due to the loss of rare species and the colonization
of pioneer species.
To test these hypotheses, we examined the shortterm (4–5 years) dynamics of stand structure and
species composition of secondary forest stands in
two chronosequences in the Luquillo and Carite
Mountains of eastern Puerto Rico, which were struck
by Hurricane Georges in 1998, 2–3 years after the
original census in 1995/1996.
2. Materials and methods
2.1. Study sites and field methods
In 1995/1996, we sampled 53 secondary forests that
developed from abandoned pastures in the Luquillo
(29 sites aged 9.5–75 years) and Carite (24 sites aged
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
4–77 years) regions. The Luquillo and Carite chronosequences occur in moist to wet subtropical forest
life-zones (Ewel and Whitmore, 1973) have similar
volcanic soils (Bocchechiamp, 1977) and have secondary forests derived from abandoned cattle pastures that are greater than 30 years old. We used aerial
photographs from 1936, 1951, 1964, 1977, 1983,
1988, and 1995 to select and date individual sites.
Site age was determined by taking the mid-point
between the last photograph with pasture and the
first photograph with shrubs or small trees. In 2000,
we resurveyed six sites in Luquillo and nine sites in
Carite of the original 53 sites, 5 and 4 years after the
original survey, respectively. These sites were chosen
to reflect a range of ages within the chronosequences
and were areas that had no signs of human intervention during the time period. Although the sites
sampled in 2000 in each chronosequences were
similar in slope, basal area, and density by age-class
in 1995/1996, they varied significantly in elevation
and species richness (Table 1). The Carite sites
occurred at a higher elevation and had higher species
richness (Table 1).
On 21 September 1998, Hurricane Georges, a
category 3 hurricane, made landfall in southeastern
Puerto Rico, and moved slowly across Puerto Rico
from east to west over the central mountain range
(Bennet and Mojica, 1998). Although winds decreased
to category 2 level immediately after landfall, most of
the island experienced sustained hurricane force
winds of 130–167 km/h and gusts of 148–209 km/h
for approximately 18 h. Rainfall totals exceeded
63.5 cm in many mountainous areas. Damage from
381
Hurricane Georges to the study sites in Luquillo and
Carite included severe defoliation and structural
damage to the canopy.
Fifteen of the 53 original sites were resampled in
May/June 2000, using the same sampling methodology as in the original survey (Aide et al., 1996;
Pascarella et al., 2000). Vegetation between 1 and
10 cm dbh was measured in four parallel 1 m 50 m
transects, and vegetation greater than 10 cm dbh was
sampled in two 10 m 50 m transects between the
four parallel transects (total area sampled ¼ 0:12 ha).
We did not sample woody vines, epiphytes, or herbaceous plants. Nomenclature follows Liogier (1985,
1988, 1994, 1995, 1997). In each site, we calculated
the basal area/ha, stem density/ha, and species richness/0.12 ha. For more details on the location of the
sites and sampling methodology, see Aide et al. (1996)
and Pascarella et al. (2000). In the recensus, any
obvious damage (e.g. top-snap or tipover) to living
stems from the hurricane was noted.
2.2. Data analysis
To determine if the original chronosequences from
each region successfully predicted the current conditions, we examined if the stands fell within the
confidence interval of the regressions from the original
chronosequences for density, basal area, and species
richness. We also examined the directional movement
of individual sites with regard to the predicted
changes in their composition. In the calculation of
the original chronosequence regression of stem density on age, three sites that had extremely high
Table 1
Regional comparison of secondary forests in Luquillo and Carite from the original sampling (Luquillo: 1995, Carite: 1996)
Variable
Elevation (m)
Slope (% grade)
Basal area (m2 ha1)
Stem density (number/ha)
Species richness (number/0.12 ha)
Luquillo (N ¼ 6)
80
31.7
23.7
8848
17.3
25.2
4.2
6.6
1129
1.8
Carite (N ¼ 9)
574.4
26.7
31.6
8086
26.0
47.4
4.3
3.5
822
2.3
Test valuea
9.22
0.79
1.80
1.57
14.50
Significance
<0.0001
NSb
NS
NS
<0.01
Data are only from sites that were resampled in 2000. N ¼ number of sites. Values are means standard errors. Tests were two-sample t-tests
for elevation and slope by region and two-way ANOVA for basal area, density, and species richness by region, age-class (<25 years, 25–45
years, >45 years), and the interaction. Only data for regional effects are presented.
a
T-statistic for t-tests and F-statistic for regional effects in two-way ANOVA. For the two-way ANOVA, age-class was significant for all
three variables and there were no significant interaction effects.
b
Not significant (P > 0:05).
382
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
densities (>15,000 stems/ha, two from Carite and one
from Luquillo) were removed from the data. These
unique sites had been pastured, abandoned, and then
recleared by hand producing many multi-stemmed
Miconia prasina shrubs. All other variables used
the entire data set from each region. Because previous
work showed that stem density increased in young
sites, decreased in intermediate sites, and stabilized in
older sites, we classified sites as young (<25 years),
medium (25–45 years), or old (>45 years). We determined if the two regions differed in rate of change in
stem density, basal area, species richness, damage to
large stems, and species turnover using ANOVA with
region and age-class as the independent variables with
Bonferroni comparisons of means (a ¼ 0:05). The
change in density, basal area, and species richness
was standardized to an annual basis because the time
between census periods was different for the two
regions. We compared the size-class distributions of
the sites between the two census dates using a Kolmogorov–Smirnov two-tailed test of equal distribution. We calculated the proportion of species lost,
gained, and lost and gained (total turnover) as the
number of lost, gained, and lost and gained species/
number of species in the original census in 1995/1996.
We classified all species that were lost from the sites in
1995/1996 by life-form and compared the proportional loss of life-forms for trees and shrubs using
ANOVA. Grouping sites by age-class, we calculated
the mean importance value of the top three species in
1995/1996 and 2000 for each region. All variables
were checked for normality. We either transformed
variables with non-normal distributions or used an
appropriate non-parametric test. All statistics were
calculated using Statistix 7 (2000).
We used a non-parametric ordination technique
(Nonmetric Multidimensional Scaling, NMS) to
examine changes in species composition along the
chronosequence. For both regions, we used the original data set (95/96) from all abandoned pastures to
create the ordination using NMS. We used the ordination that best minimized the stress and the number of
dimensions (Luquillo: three-dimensional, Carite: twodimensional). For both regions, we used the Sørenson
distance measure. We then used NMS scores to determine if the original multivariate ordination of each
region successfully predicted the location of the
2000 census sites (McCune and Medford, 1999).
The procedure NMS scores fitted the new points
(the resampled sites in 2000) into the original ordination space and examined the stress of the fit and the
placement of the new points within the expected
ordination space. To determine how site age affected
the rate of change in species composition, we compared the Sørenson dissimilarity values between
sampling periods within each site. PC-ORD Version
4 was used for all ordinations (McCune and Medford,
1999).
3. Results
3.1. Stand structure–stem density and size-class
distribution
There was a large decrease in stem density in all
sites, with an average decrease of 47% (range 22–
61%) (Fig. 1A and B, Appendix A). There was a
significant effect of age-class on stem density change
(F½2;14 ¼ 8:9, P < 0:01), but there was no effect of
region or the interaction. All age-classes showed a
significant decrease in stem density, but intermediate
sites had a significantly greater decrease than young or
older sites (Mean S:D: decrease in stems:
young ¼ 619:63 271:12 (n ¼ 4), intermediate ¼
1318:8 208:31 (n ¼ 5), and old ¼ 772:67 383:44 (n ¼ 6)). In Luquillo, none of the resampled
sites fell within the confidence intervals of the chronosequence derived from the original surveys for stem
density (Fig. 1A). At Carite, only the oldest four sites
fell within the confidence intervals of the original
chronosequence (Fig. 1B). All the sites showed a
lower than expected stem density, even for sites where
the chronosequence predicted a strong decline. Examination of the residuals revealed that sites in both
regions had higher densities than predicted in 1995/
1996 but much lower than predicted stem densities in
2000 (589 versus 4881 and 963 versus 3107 for
Luquillo and Carite in 1995/1996 and 2000, respectively).
At the stand level, there was significant change in the
size-class distribution of stems across all age-classes,
although the smallest change occurred in the young and
older sites (Fig. 2, Appendix A). Across all sites, there
was a large decline in the number of stems between 1
and 5 cm in dbh and a smaller decline in stems between
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
383
Fig. 1. Predicted and actual density (A, B), basal area (C, D), and species richness (E, F) of abandoned pastures in Luquillo (A, C, E) and
Carite (B, D, F). Dashed lines indicate the 95% confidence interval of the regression from the 1995/1996 samples of the original census; sites
in 1995 are in solid circles; open circles are from resampled sites in 2000; lines connect the same sites. Density and basal area are calculated
per hectare while species richness is from 0.12 ha.
5 and 10 cm dbh. At most sites, there was a large
increase in stems between 10 and 15 cm. Size-classes
larger than 15 cm showed variable responses across
sites, with some increasing and others decreasing. At
the species level, the tree Tabebuia heterophylla showed
a decrease in stems <10 cm but a large increase in stems
>10 cm, particularly in the intermediate aged sites. In
contrast, the shrubs M. prasina and Casearia sylvestris
had a large decrease in small stems across all sites
(Fig. 3).
3.2. Stand structure–basal area
Basal area decreased in 11 of the 15 sites (5–45%
decrease) and increased in four sites (2–25%
increase), with an average decrease of 11% across
all sites in both regions (Fig. 1C and D, Appendix A).
There was no significant age-class, regional, or interaction effect on change in basal area/year. Change in
basal area/year was not correlated with the percent of
damaged stems >10 cm dbh. At Luquillo, only one of
384
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
Fig. 2. Size class distributions at Luquillo and Carite in 1995/1996 and 2000. Sites are arranged vertically from young to older sites (one site
per panel for sites that had two replicates).
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
385
Fig. 3. The number of individuals by age-class pre-and post-hurricane for the tree (A, B) and shrub species (C, D) in each region that had the
greatest change in average importance value post-hurricane. For T. heterophylla, the size class distribution is also shown (<10 cm dbh, >10 cm
dbh). For the two shrubs, the majority of stems were <10 cm.
the six resampled sites fell within the confidence
intervals of the chronosequence derived from the
original surveys for basal area (Fig. 1C). At Carite,
six of the nine sites fell within the confidence intervals
of the original chronosequence (Fig. 1D). Sites from
Luquillo tracked the chronosequence slightly better,
with two of the six sites showing an increase while
only two of nine sites at Carite increased. The two
largest decreases in basal area occurred at Carite.
Examination of the residuals revealed that sites in
both regions were relatively close to the predicted
values in 1995/1996, but had much lower basal areas
than predicted in 2000 (0.72 versus 7.16 and 1.29
versus 5.79 for Luquillo and Carite in 1995/1996
and 2000, respectively).
3.3. Species richness
Across all sites in both regions, species richness
declined in 11 sites, three stayed the same, and one site
increased in species richness for an average loss of
16% (Fig. 1E and F, Appendix A). There was a
significant effect of both region (F½1;14 ¼ 12:5,
P < 0:01) and age-class (F½2;14 ¼ 11:1, P < 0:01)
but not their interaction. Sites in Carite had a significantly greater mean loss in species richness compared with sites in Luquillo (Mean S:D: ¼
1:33 0:37, n ¼ 9; 0:43 0:23 species, n ¼ 6,
respectively) and sites in the 25–45-year age-class had
a significantly greater mean loss in species richness
than younger or older sites (Mean S:D: ¼ 1:91 386
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
1:02, 0:48 0:71, and 0:53 0:63, for intermediate (n ¼ 5), young (n ¼ 4) and old sites (n ¼ 6),
respectively).
At Luquillo, four of the six resampled sites (67%),
with the exception of older sites, fell within the
confidence intervals of the chronosequence derived
from the original surveys for species richness while
the other two were very close to the predicted range of
values (Fig. 1E). At Carite, five of the nine sites (56%)
fell within the confidence intervals of the original
chronosequence, with much lower than expected
species richness in younger sites (Fig. 1F). In both
regions, no sites had higher than expected species
richness with most sites falling near the lower confidence interval estimate. Examination of the residuals revealed that sites in both regions were
relatively close to the predicted values in 1995/
1996 but had much lower species richness than predicted in 2000 (0.74 versus 2.86 and 1.70 versus
4.46 for Luquillo and Carite in 1995/1996 and 2000,
respectively).
All sites gained and lost species during this period,
with sites in Luquillo losing 2–10 species and gaining
1–4 species and sites in Carite losing from 3 to 14
species and gaining from 1 to 6 species (Appendix A).
There was a significant effect of both age-class and
region on number of species lost (F½2;14 ¼ 14:8,
P < 0:01; F½1;14 ¼ 14:8, P < 0:01, for age-class and
region, respectively) and total species turnover
(F½2;14 ¼ 8:8, P < 0:01; F½1;14 ¼ 10:3, P < 0:05,
respectively) but not the number of species gained.
For species lost and total turnover, intermediate ageclasses had significantly higher loss than young or
older sites. There was no effect of age-class, region, or
their interaction on the proportion of species gained.
There was a significant effect of age-class but not
region nor their interaction on the proportion of species lost (F½2;14 ¼ 12:3, P < 0:01). Sites aged 25–45
had the highest proportion of species lost
(45:2 7:2%) while young sites were intermediate
(31:0 14%) and older sites had the lowest proportion
of species lost (21:2 6:8%). Young sites (<25 years)
had the highest proportional gain (20% versus 13–
14% for older sites). There was a significant effect of
age-class ((F½2;14 ¼ 9:7, P < 0:01) but not region or
interaction on proportional total turnover, with young
and intermediate sites having higher proportional
turnover than older sites (51:5 16:7, 58:8 8:7,
and 47:1 15:2%, for young (n ¼ 4), intermediate
(n ¼ 5), and older sites (n ¼ 6), respectively).
Classified by life-form, ANOVA found that there
was no effect of region, age-class, or their interaction
on the proportion of tree species lost, but a significant
effect of both age-class and region on the proportion of
shrub species lost. Carite suffered higher proportional
loss of shrubs than Luquillo (F½1;14 ¼ 10:9, P < 0:01;
43 5% versus 28 8%) and intermediate sites suffered higher proportional loss than young or old sites
(F½2;14 ¼ 10:2, P < 0:01; 52 4% for intermediate,
24 13% for young, and 32 3% for old sites).
3.4. Species composition
The important values of species that were lost
during the period ranged from 0.8 to 27.2% in
Luquillo and from 1.7 to 33.8% in Carite (Appendix
A). The importance value of species gained during the
period ranged from 0.9 to 9.2% in Luquillo and from
0.5 to 17.2% in Carite (Appendix A). The mean
importance value of the species that were involved
in the turnover (average importance value of all species lost and gained from the stand at the two censuses)
ranged from 1.5 to 18.2% in Luquillo and from 3.3 to
20.9% in Carite (Appendix A). There was no significant difference between the two regions, age-classes,
or their interaction in importance value of species lost,
gained, or the mean importance value of the turnover
species. Both regions combined had a significant
negative correlation of importance value of turnover
species with stand age (r ¼ 0:55, P < 0:05,
d:f: ¼ 14) (Fig. 4A). At the species level, most of
the abundant species remained abundant, although
there was a shift in the rankings, particularly in the
youngest sites (Table 2). For example, the two Casearia species tended to decline, while T. heterophylla
showed a large increase.
The NMS scores procedure that used the original
ordination of the pasture sites from Luquillo and
Carite successfully predicted the ordination location
of the resampled sites in 2000 with acceptable stress
and within the expected axis score space (Luquillo
critical stress ¼ 13:93, Carite critical stress ¼ 13:78).
The Sørenson dissimilarity index, comparing the
species composition between sampling periods within
each site, was negatively correlated with site age
(Fig. 4B). Older sites had a lower dissimilarity index,
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
387
Fig. 4. (A) Mean importance value of turnover species (species lost or gained from a site) vs. site age. (B) Dissimilarity index (Sørensen)
within a site vs. site age. For both A and B, the regression is from Luquillo and Carite combined.
and thus there was little change in the species composition.
4. Discussion
Between the two censuses, secondary forests in
both regions experienced considerable dynamics.
How well did the chronosequences predict the
dynamics? In general, the chronosequence model
was relatively accurate predicting changes in basal
area and species composition, but the dramatic
decrease in stem density and species richness was
not predicted by the chronosequence. We attribute this
deviation from the expected results directly to the
effects of Hurricane Georges.
Although some sites should have naturally shown a
decrease in density due to successional changes as
predicted by the chronosequence (Fig. 1), the large
decrease across sites of all ages was due to the damage
388
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
Table 2
The species in both census intervals that had the highest three
importance values, grouped by age-class and region
Region
Age
Species
IV
(1995/1996)
IV
(2000)
Carite
<25
S. campanulata
31.6
C. sylvestris
20.5
T. heterophylla
7.9
Casearia guianensis 8.7
Myrcia splendens
6.1
32.6
5.9
42.8
5.7
17.0
Carite
25–45
S. campanulata
T. heterophylla
Guarea guidonia
47.0
16.7
10.6
43.5
22.3
17.2
Carite
>45
Myrcia deflexa
Prestoea montana
Syzygium jambos
Casearia arborea
13.1
10.8
8.9
10.1
10.5
9.3
9.3
7.6
Luquillo
<25
M. prasina
C. guianensis
Didymonopanax
morototoni
M. splendens
T. heterophylla
30.2
18.6
16.0
24.8
10.5
0
3.2
2.1
15.3
13.3
Luquillo
25–45
T. heterophylla
C. guianensis
M. prasina
Ocotea leucoxylon
32.9
11.7
11.4
3.9
48.0
11.8
6.0
9.0
Luquillo
>45
Mangifera indica
G. guidonia
Psychotria brachiata
O. leucoxylon
26.7
18.8
16.7
14.7
24.3
14.9
20.9
10.1
More than three species in a group indicates that the top three
species changed during the time period.
of Hurricane Georges. Hurricane winds directly
affected large individuals by uprooting and bole snapping, and indirectly affected understory individuals by
falling canopy trees and debris (Frangi and Lugo,
1991; Basnet et al., 1992; Zimmerman et al., 1994;
Fu et al., 1996). Although understory individuals are
relatively protected from the direct winds of the
hurricane, there was high mortality in the small size
classes, apparently due to the large amount of falling
debris. Limbfalls and other falling debris from the
canopy are important causes of mortality in nonhurricane prone areas (Aide, 1987), so it is likely that
this is a major source of mortality during hurricane
events.A secondary effect of the dramatic decrease in
density was the decrease in species richness. In tro-
pical forests, there is a strong effect of density on
species richness due to the patchy and low-density
distribution of many tropical plant species (Denslow,
1995). The reduction in stem density that occurred in
the two regions was primarily in the small size classes,
affecting many shrub and saplings that were not well
represented in the subcanopy or canopy. In many sites,
the loss of these stems resulted in the loss of individual
species, and the loss of species was not compensated
by the recruitment of new species (e.g., early successional or pioneer species). The lack of substantial
pioneer species recruitment into hurricane-damaged
secondary forest sites (Zimmerman et al., 1995) is
related to the relative scarcity of tip-ups mounds, pits,
and large treefall gaps. This may be an important
limitation for pioneer species recruitment in young
secondary forests (Pascarella, 1997; Brokaw, 1998).
Similarly, Lomáscolo and Aide (2001) and Imbert
et al. (1998) have documented that most recruitment
into sapling size classes in forests following hurricane
damage originates from individuals in the seedling
bank, which established prior to the hurricane. These
studies suggest that changes in species richness will
depend more on the composition of the seedling bank,
rather than recruitment of pioneer species immediately following the hurricane.
The lower species richness documented in this
study may be a temporary result of the hurricane.
Because we only measured stems 1.3 m in height
and 1 cm in diameter, damaged stems that were alive
and had resprouted but had not yet entered into these
classes were not measured. Some species that were
‘‘lost’’ from the sites are probably still present as small
individuals. Other studies have also documented a
temporary reduction in species richness followed by
an increase in species 5 years post-hurricane (Fu et al.,
1996); however, Imbert et al. (1998) caution that
frequent strong hurricanes may lead to the extinction
of rare, non-pioneer species. Most of the species that
were lost from our sites were rare, non-pioneer species, although some pioneer species were also lost in
young sites.
In comparison to stem density and species richness,
the chronosequence prediction of change in basal area
was only slightly affected by Hurricane Georges. The
overall change in basal area integrated both the negative effects of the disturbance, the death of individual
stems, and the positive effects of reduced competition
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
and increased light and nutrient availability, resulting
in increased growth of surviving stems into larger size
classes. Although many small plants were killed by the
hurricane, these individuals do not contribute much to
the total basal area. More important was the large
increase in the number of stems in the 10–20 cm size
class that make a significant positive contribution to
basal area. Our results are similar to those found by
Scatena et al. (1996) for two mature forest sites
following Hurricane Hugo. In their study, higher net
primary productivity returned the sites to 86% prehurricane values within 5 years. Only two sites had a
large decrease in basal area and they were dominated
(56 and 47% importance value, respectively) by large
diameter individuals of the introduced tree Spathodea
campanulata. S. campanulata is a short-lived, low
wood-density tree that is very susceptible to high
winds (Lomáscolo and Aide, 2001).
Although species richness was strongly reduced, the
species composition showed little change because the
species that were lost or gained at a site were of little
relative importance. The ordination used the relative
importance value of the species as its quantitative
estimate. Importance values in this study were the
average relative basal area and relative abundance of
each species in each site. The majority of the species
gains and losses were species with few stems, small
diameters or both, thus having little effect on stand
composition. Although there was little overall change,
there was substantially more change in the younger
stands than in the older stands, as predicted by the
original NMS ordination of the chronosequence and
by our previous studies (Aide et al., 1995, 2000;
Pascarella et al., 2000). This pattern is common across
most successional studies in a variety of habitats
(Baker et al., 1996; Debussche et al., 1996; Foster
and Tilman, 2000) due to increased resistance to
change with stand age. Other studies of hurricane
impacts have found similar trends, with similar minor
changes in species composition across secondary forests of different ages or before-and-after hurricane
disturbance at a given site (Zimmerman et al., 1994;
Fu et al., 1996; Imbert et al., 1998; Lomáscolo and
Aide, 2001).
In spite of the limited compositional change, there
was a high degree of species turnover at the local site
level. At the regional level, the change was much less,
indicating that the loss and gain of species is primarily
389
local and not regional. Regional diversity is likely
more influenced by the spatial dynamics of land-use
than the more transient effects of hurricane disturbance. In both regions, secondary forests occur in a
complex mosaic of regenerating forest, agriculture,
and urban and suburban developments (Thomlinson
et al., 1996; Foster et al., 1999; Pascarella et al., 2000),
creating differences in the relative abundance of
natives versus exotics, availability and abundance of
dispersal agents, and distance to seed sources. Regional diversity may also be influenced by environmental
differences between the two sites, particularly the
higher elevations found in the Carite study region.
4.1. Relative effects of hurricanes and land-use
history on secondary succession
The interaction between humans and natural disturbances is a fundamental concept in understanding
forest dynamics. In both temperate and tropical areas,
hurricanes have been implicated as important natural
disturbances (Foster and Boose, 1992; Boose et al.,
1994; Zimmerman et al., 1996). Foster et al. (1999)
studied the landscape-scale dynamics of the Luquillo
Experimental Forest and emphasized the importance
of the interaction between natural disturbance (e.g.,
hurricanes) and land-use history on forest ecosystems.
In this study, the impact of hurricanes did not substantially alter the species composition or successional
dynamics of secondary forests. Although plant density
was dramatically reduced, and some shrub and rare
species may have been lost at a local scale, there was
little effect on basal area and species composition.
Three previous chronosequence studies in Puerto
Rico (Aide et al., 1996; Zimmerman et al., 1995;
Lomáscolo and Aide, 2001) have examined the relative roles of hurricanes and land-use history in forest
structure and composition. They found that hurricanes
had a minor effect compared with the anthropogenic
effect of land-use history (uncut forest, used as pasture, or abandoned coffee plantation). The major
differences between these studies and the observations
of Foster et al. (1999) are the age and structure of the
forest. Foster et al. (1999) studied mature forest with a
much greater diversity of forest structure, which is
more vulnerable to hurricane damage. Furthermore, in
mature forest, the presence of larger canopy trees
results in larger tree falls, which can favor the colo-
390
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
nization of early successional species. In contrast, the
vertical structure in secondary forest is more simple
and homogeneous, and there are fewer large canopy
trees. Our results also support the observations of Fu
et al. (1996), who found similar patterns in a paired
site of a plantation and adjacent secondary forest, with
land-use effects overriding any effects of hurricane
disturbance. The occurrence of hurricanes during
secondary succession will increase the variation in
secondary forest structure, but should not substantially
alter the overall successional process or species composition. However, as these secondary forests age, we
predict that the effect of hurricanes on forest structure
and ecosystem process will become more important,
reflecting the increased structural diversity found in
older forest stands (Zimmerman et al., 1996; Foster
et al., 1999; Lomáscolo and Aide, 2001).
Acknowledgements
This research was supported by funds from NASAIRA to the University of Puerto Rico-Rio Piedras. We
thank the landowners in the two areas for permission
to use their lands for this study.
Appendix A
Region
and site
no.a
Ageb Basal
area
(m2 ha1)
1995/1996
Basal area
(m2 ha1)
2000
Density
(stems ha1)
1995/1996
Density
(stems ha1)
2000
%
Damaged
stems
Species
K–S test
richness
of dbh
distributionc (number/0.12 ha)
95/96
Species
richness
(number/0.12 ha)
2000
Species
lost
Species
gained
Total
species
turnover
IV of
losses
(%)
IV of
gains
(%)
Mean IVd
of turnover
sp. (%)
L-10
L-12
L-7
L-8
L-1
L-9
C-6
C-8
C-10
C-12
C-16
C-19
C-20
C-21
C-23
14.5
20
30.5
42.5
56.5
80
17
17
29
29
41
56
56
81
81
4.0
8.6
30.4
15.7
33.6
36.2
16.9
22.4
29.6
20.0
30.7
30.6
24.9
21.8
39.8
4940
8850
12580
11210
7030
8480
5170
6670
9670
8330
12520
9220
7430
9260
4500
3860
4850
6070
4790
3190
3840
2210
3780
5362
3240
5890
4690
5170
4780
3070
0
0
9
7
13
5
7
2
23
0
3
0
15
5
3
NS
8
15
16
17
19
16
11
19
11
21
19
29
23
31
21
4
2
7
10
2
5
3
11
13
14
13
6
9
7
4
2
2
4
4
4
1
3
5
1
6
1
1
4
5
4
6
4
11
14
6
6
6
16
14
20
14
7
13
12
8
27.2
2.6
12.7
17.8
0.8
3.6
1.7
33.8
9.6
13.4
10.7
6.0
5.0
6.5
2.0
9.2
3.5
3.3
6.2
2.2
0.9
5.2
7.9
1.1
17.2
0.5
0.7
3.7
4.5
5.0
18.2
3.1
8.0
12.0
1.5
2.3
3.5
20.9
5.4
15.3
5.6
3.3
4.4
5.5
3.5
a
4.8
9.0
34.7
15.2
33.0
45.4
30.1
17.9
53.8
22.0
37.6
32.2
27.7
27.3
35.7
****
****
****
****
**
NS
*
NS
****
****
***
NS
***
NS
10
15
19
23
17
20
11
25
23
29
31
33
28
33
21
L ¼ Luquillo, C ¼ Carite. Site numbers refer to appendices in Aide et al. (1996) and Pascarella et al. (2000).
Estimated age in recensus of 2000. Sites were originally sampled in 1995 (Luquillo) and 1996 (Carite).
c
NS ¼ not significant.
d
Importance value. The sum of the importance values of the species that were lost and gained during the period.
*
P < 0:05.
**
P < 0:01.
***
P < 0:001.
****
P < 0:0001.
b
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
Characteristics and dynamics of secondary forests in Luquillo and Carite from 1995 to 2000 for basal area, stem density, percent damaged
stems, species richness, species loss, species gain, and importance value of turnover species. Data are from 1995 (Luquillo), 1996 (Carite), and
2000 (both regions).
391
392
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
References
Aide, T.M., 1987. Limbfalls: a major cause of sapling mortality for
tropical forest plants. Biotropica 19, 284–285.
Aide, T.M., Zimmerman, J.K., Herrera, L., Rosario, M., Serrano,
M., 1995. Forest recovery in abandoned tropical pastures in
Puerto Rico. For. Ecol. Manage. 77, 77–86.
Aide, T.M., Zimmerman, J.K., Rosario, M., Marcano, H., 1996.
Forest recovery in abandoned cattle pastures along an
elevational gradient in northeastern Puerto Rico. Biotropica
28, 537–548.
Aide, T.M., Zimmerman, J.K., Pascarella, J.B., Rivera, L.,
Marcano, H., 2000. Forest regeneration in a chronosequence
of tropical abandoned pastures: implications for restoration
ecology. Restor. Ecol. 8, 328–338.
Aplet, G.H., Vitousek, P.M., 1994. An age altitude matrix analysis
of Hawaiian rain-forest succession. J. Ecol. 82, 137–147.
Baker, J.P., Olff, H., Willems, J.H., Zobel, M., 1996. Why do we
need permanent plots in the study of long-term vegetation
dynamics. J. Veg. Sci. 7, 147–155.
Basnet, K., Likens, G.E., Scatena, F.N., Lugo, A.E., 1992.
Hurricane Hugo: damage to a tropical rain forest in Puerto
Rico. J. Trop. Ecol. 8, 47–55.
Bennet, S.P., Mojica, R., 1998. Hurricane Georges Preliminary
storm report: from the tropical Atlantic to the United States
Virgin Islands and Puerto Rico. A NOAA National Weather
Service NWSFO San Juan Official Report, October 1998,
pp. 1–15.
Birdsey, R.A., Weaver, P.L., 1987. Forest area trends in Puerto
Rico. United States Forest Service Research Report Note SO331. Southern Forest Experimental Station, New Orleans, LA,
USA.
Boccheciamp, R.A., 1977. Soil survey of the Humacao area of
Eastern Puerto Rico. United States Department of Agriculture
Soil Conservation Service, San Juan, PR, USA.
Boose, E.R., Foster, D.R., Fluet, M., 1994. Hurricane impacts to
tropical and temperate landscapes. Ecol. Monogr. 64, 369–400.
Brokaw, N.V.L., Grear, J.S., 1991. Forest structure before and after
Hurricane Hugo at three elevations in the Luquillo mountains.
Puerto Rico Biotropica 23, 386–392.
Brokaw, N.V.L., Walker, L., 1991. Summary of the effects of
Caribbean Hurricanes on vegetation. Biotropica 23, 442–447.
Brokaw, N.V.L., 1998. Cecropia schreberiana in the Luquillo
Mountains of Puerto Rico. Bot. Rev. 64, 91–120.
Collins, S.L., Adams, D.E., 1983. Succession in grasslands: thirtytwo years of change in a central Oklahoma tallgrass prairie.
Vegetatio 51, 181–190.
Crow, T.R., 1980. A rainforest chronicle: a 30-year record of
change in structure and composition at El Verde, Puerto Rico.
Biotropica 12, 45–55.
Davis, A.L.V., van Aarde, R.J., Scholtz, C.H., Delport, J.H., 2003.
Convergence between dung beetle assemblages of a postmining vegetational chronosequence and unmined dune forest.
Restor. Ecol. 11, 29–42.
Debussche, M., Escarré, J., Lepart, J., Houssard, C., Lavorel, S.,
1996. Changes in Mediterranean plant succession: old-fields
revisited. J. Veg. Sci. 7, 519–526.
Denslow, J.S., 1995. Disturbance and diversity in tropical rain
forests: the density effect. Ecol. Appl. 5, 962–968.
Dewalt, S.J., Schnitzer, S.A., Denslow, J.S., 2000. Density and
diversity of lianas along a chronosequence in a central
Panamanian lowland forest. J. Trop. Ecol. 16, 1–19.
Dietz, J.L., 1986. Economic History of Puerto Rico. Princeton
University Press, Princeton, NJ, USA.
Erickson, D.L., Hamrick, J.L., 2003. Genetic and clonal diversity
for Myrica cerifera along a spatiotemporal island chronosequence. Heredity 90, 25–32.
Ewel, J.J., Whitmore, J.L., 1973. The ecological life-zones of
Puerto Rico and the US Virgin Islands. US Forest Service
Research Paper ITF-18, Rio Piedras, PR, USA.
Fastie, C.L., 1995. Causes and ecosystem consequences of multiple
pathways of primary succession at Glacier Bay. Alaska Ecol.
76, 1899–1916.
Fernandez, D.S., Fetcher, N., 1991. Changes in light availability
following Hurricane Hugo in a subtropical Montane forest in
Puerto Rico. Biotropica 23, 393–399.
Foster, B.L., Tilman, D., 2000. Dynamic and static views of
succession: testing the descriptive power of the chronosequence
approach. Plant Ecol. 146, 1–10.
Foster, D.R., Boose, E.R., 1992. Species and stand response
to catastrophic wind in New England. USA. J. Ecol. 76, 135–
151.
Foster, D.R., Fluet, M., Boose, E.R., 1999. Human or natural
disturbance: landscape-scale dynamics of the tropical forests of
Puerto Rico. Ecol. Appl. 9, 555–572.
Frangi, J.L., Lugo, A.E., 1991. Hurricane damage to a flood plain
forest in the Luquillo mountains of Puerto Rico. Biotropica 23,
324–335.
Fu, S., Pedraza, C.R., Lugo, A.E., 1996. A twelve-year comparison
of stand changes in a Mahogany plantation and a paired natural
forest of similar age. Biotropica 28, 515–524.
Gerrish, G., Mueller-Dombois, D., 1999. Measuring stem growth
rates for determining age and cohort analysis of a tropical
evergreen tree. Pac. Sci. 53, 418–429.
Grau, H.R., Arturi, M.F., Brown, A.D., Acenolaza, P.G., 1997.
Floristic and structural patterns along a chronosequence of
secondary forest succession in Argentinean subtropical montane forests. For. Ecol. Manage. 95, 161–171.
Imbert, D., Rousteau, A., Labbe, P., 1998. Hurricanes and
biological diversity in tropical forests—the case of Guadeloupe.
Acta Oecol. Int. J. Ecol. 19, 251–262.
Kennard, D.K., 2002. Secondary forest succession in a tropical dry
forest: patterns of development across a 50-year chronosequence in lowland Bolivia. J. Trop. Ecol. 18, 53–66.
Lehmkuhl, J.F., Everett, R.L., Schellhaas, R., Ohlson, P., Keenum,
D., Riesterer, H., Spurbeck, D., 2003. Cavities in Snags along a
Wildfire chronosequence in Eastern Washington. J. Wildlife
Manage. 67, 219–228.
Liogier, A.H., 1985. Descriptive Flora of Puerto Rico and Adjacent
Islands. Spermatophyta. Vol. I: Casuarinaceae to Connaraceae.
Editorial de la Universidad de Puerto Rico, Rio Piedras, PR,
USA.
Liogier, A.H., 1988. Descriptive Flora of Puerto Rico and Adjacent
Islands. Spermatophyta. Vol. II: Leguminosae to Anacardia-
J.B. Pascarella et al. / Forest Ecology and Management 199 (2004) 379–393
ceae. Editorial de la Universidad de Puerto Rico, Rio Piedras,
PR, USA.
Liogier, A.H., 1994. Descriptive Flora of Puerto Rico and Adjacent
Islands. Spermatophyta. Vol. III: Cyrillaceae to Myrtaceae.
Editorial de la Universidad de Puerto Rico, Rio Piedras, PR,
USA.
Liogier, A.H., 1995. Descriptive Flora of Puerto Rico and Adjacent
Islands. Spermatophyta. Vol. IV: Melastomataceae to Lentibulariaceae. Editorial de la Universidad de Puerto Rico, Rio
Piedras, PR, USA.
Liogier, A.H., 1997. Descriptive Flora of Puerto Rico and Adjacent
Islands. Spermatophyta. Vol. V: Acanthaceae to Compositae.
Editorial de la Universidad de Puerto Rico, Rio Piedras, PR,
USA.
Lodge, D.J., Scatena, F.N., Asbury, C.E., Sanchez, M.J., 1991. Fine
litterfall and related nutrient inputs resulting from Hurricane
Hugo in subtropical wet and lower Montane rain forests of
Puerto Rico. Biotropica 23, 336–342.
Lodge, D.J., McDowell, W.H., 1991. Summary of ecosystem-level
effects of Caribbean Hurricanes. Biotropica 23, 373–378.
Lomáscolo, T., Aide, T.M., 2001. Seed and seedling bank dynamics
in secondary forests following Hurricane Georges in Puerto
Rico. Car. J. Sci. 37, 259–270.
Lugo, A.E., Ramos, O., Molina, S., Scatena, F.N., VélezRodrı́guez, L.L., 1996. A fifty-three year record of land use
change in the Guanica forest biosphere reserve and its vicinity.
USDA Forest Service International Institute of Tropical
Forestry, Rio Piedras, PR, USA.
Marcano-Vega, H., Aide, T.M., Baez, D., 2002. Forest regeneration
in abandoned coffee plantations and pastures in the Cordillera
Central of Puerto Rico. Plant Ecol. 161, 75–87.
McCune, B., Medford, M.J., 1999. PC-ORD for Windows.
Multivariate analysis of ecological data, Version 4.10. MjM
Software Design, Gleneden Beach, OR, USA.
Pascarella, J.B., 1997. Hurricane disturbance and the regeneration
of Lysiloma latisiliquum (Fabaceae): a tropical tree in south
Florida. For. Ecol. Manage. 92, 97–106.
Pascarella, J.B., Aide, T.M., Serrano, M.I., Zimmerman, J.K., 2000.
Land-use history and forest regeneration in the Cayey
Mountains, Puerto Rico. Ecosystems 3, 217–228.
Pickett, S.T.A., 1989. Space for time substitution as an alternative
to long-term studies. In: Likens, G.E. (Ed.), Long-term Studies
in Ecology. Wiley, Chicester, pp. 71–88.
393
Rivera, L.W., Aide, T.M., 1998. Forest recovery in the Karst region
of Puerto Rico. For. Ecol. Manage. 108, 63–75.
Rivera, L.W., Zimmerman, J.K., Aide, T.M., 2000. Vegetation
analysis of secondary forests in the Karst Region of the
Dominican Republic. Plant Ecol. 148, 115–125.
Rudel, T.K., Perez-Lugo, M., Zichal, H., 2000. When fields revert
to forest: development and spontaneous reforestation in postwar Puerto Rico. Prof. Geog. 52, 386–397.
Saldariagga, J.C., West, D.C., Tharp, M.L., Uhl, C., 1988. Longterm chronosequence of forest succession in the upper Rio
Negro of Colombia and Venezuela. J. Ecol. 76, 938–958.
Scatena, F.N., Moya, S., Estrada, C., Chinea, J.D., 1996. The first
five years in the reorganization of aboveground biomass and
nutrient use following Hurricane Hugo in the Bisley experimental watersheds, Luquillo experimental forest, Puerto Rico.
Biotropica 28, 424–440.
Statistix 7, 2000. Analytical Software, Tallahassee, FL, USA.
Tanner, E.V.J., Kapos, V., Healey, J.R., 1991. Hurricane effects on
forest ecosystems in the Caribbean. Biotropica 23, 513–521.
Thomlinson, J.R., Serrano, M.I., Lopezdel, T.M., Aide, T.M.,
Zimmerman, J.K., 1996. Land-use dynamics in a postagricultural Puerto Rican landscape (1936–1988). Biotropica
28, 525–536.
Twigg, L.E., Fox, B.J., Jia, L., 1989. The modified primary
succession following sand mining: a validation of the use of
chronosequence analysis. Austral. J. Ecol. 14, 441–447.
Verchot, L.V., Davidson, E.A., Cattanio, J.H., Ackerman, I.L., 2000.
Land-use change and biogeochemical controls of methane fluxes
in soils of eastern amazonia. Ecosystems 3, 41–56.
Waide, R.B., 1991. Summary of the response of animal populations
to Hurricanes in the Caribbean. Biotropica 23, 508–512.
Walker, L.R., 1991. Tree damage and recovery from Hurricane
Hugo in the Luquillo Experimental Forest. Puerto Rico
Biotropica 23, 379–385.
Zimmerman, J.K., Everham III, E.M., Waide, R.B., Lodge, D.J.,
Taylor, C.M., Brokaw, N.V.L., 1994. J. Ecol. 82, 911–922.
Zimmerman, J.K., Aide, T.M., Rosario, M., Serrano, M., Herrera,
L., 1995. Effects of land management and a recent hurricane on
forest structure and composition in the Luquillo Experimental
Forest, Puerto Rico. For. Ecol. Manage. 77, 65–76.
Zimmerman, J.K., Willig, M.R., Walker, L.R., Silver, W.L., 1996.
Introduction: disturbance and Caribbean ecosystems. Biotropica 28, 414–423.