ArticlePDF Available

Seed rain under tree islands planted to restore degraded land in a tropical agricultural landscape

Authors:

Abstract and Figures

Planting native tree seedlings is the predominant restoration strategy for accelerating forest succession on degraded lands. Planting tree "islands" is less costly and labor intensive than establishing larger plantations and simulates the nucleation process of succession. Assessing the role of island size in attracting seed dispersers, the potential of islands to expand through enhanced seed deposition, and the effect of planting arrangements on seed dispersal by birds and bats informs restoration design. Determining the relative importance of local restoration approach vs. landscape-level factors (amount of surrounding forest cover) helps prioritize methods and locations for restoration. We tested how three restoration approaches affect the arrival of forest seeds at 11 experimental sites spread across a gradient of surrounding forest cover in a 100-km2 area of southern Costa Rica. Each site had three 50 x 50 m treatments: (1) control (natural regeneration), (2) island (planting tree seedlings in patches of three sizes: 16 m2, 64 m2, and 144 m2), and (3) plantation (planting entire area). Four tree species were used in planting (Terminalia amazonia, Vochysia guatemalensis, Erythrina poeppigiana, and Inga edulis). Seed rain was measured for 18 months beginning approximately 2 years after planting. Plantations received the most zoochorous tree seeds (266.1 +/- 64.5 seeds x m(-2) x yr(-1) [mean +/- SE]), islands were intermediate (210.4 +/- 52.7 seeds x m(-2) x yr(-1)), and controls were lowest (87.1 +/- 13.9 seeds x m(-2) x yr(-1)). Greater tree seed deposition in the plantations was due to birds (0.51 +/- 0.18 seeds x m(-2) x d(-1)), not bats (0.07 +/- 0.03 seeds x m(-2) x d(-1)). Seed rain was primarily small-seeded, early-successional species. Large and medium islands received twice as many zoochorous tree seeds as small islands and areas away from island edges, suggesting there is a minimum island size necessary to increase seed deposition and that seed rain outside of planted areas is strongly reduced. Planting design was more important for seed deposition than amount of forest cover within the surrounding 100- and 500-m radius areas. Establishing plantations and large islands facilitates the arrival of early-successional tree seeds and represents a broadly applicable strategy for increasing seed rain on abandoned agricultural lands. However, more intensive restoration approaches may be necessary for establishment of dispersal-limited species.
Content may be subject to copyright.
Ecological Applications, 20(5), 2010, pp. 1255– 1269
Ó2010 by the Ecological Society of America
Seed rain under tree islands planted to restore degraded lands
in a tropical agricultural landscape
R. J. COLE,
1,3
K. D. HOLL,
1
AND R. A. ZAHAWI
2
1
Environmental Studies Department, University of California, Santa Cruz, California 95064 USA
2
Las Cruces Biological Station, Organization for Tropical Studies, San Vito, Costa Rica
Abstract. Planting native tree seedlings is the predominant restoration strategy for
accelerating forest succession on degraded lands. Planting tree ‘‘islands’’ is less costly and
labor intensive than establishing larger plantations and simulates the nucleation process of
succession. Assessing the role of island size in attracting seed dispersers, the potential of
islands to expand through enhanced seed deposition, and the effect of planting arrangements
on seed dispersal by birds and bats informs restoration design. Determining the relative
importance of local restoration approach vs. landscape-level factors (amount of surrounding
forest cover) helps prioritize methods and locations for restoration. We tested how three
restoration approaches affect the arrival of forest seeds at 11 experimental sites spread across a
gradient of surrounding forest cover in a 100-km
2
area of southern Costa Rica. Each site had
three 50 350 m treatments: (1) control (natural regeneration), (2) island ( planting tree
seedlings in patches of three sizes: 16 m
2
,64m
2
, and 144 m
2
), and (3) plantation (planting
entire area). Four tree species were used in planting (Terminalia amazonia,Vochysia
guatemalensis,Erythrina poeppigiana, and Inga edulis). Seed rain was measured for 18 months
beginning ;2 years after planting.
Plantations received the most zoochorous tree seeds (266.1 664.5 seedsm
2
yr
1
[mean 6
SE]), islands were intermediate (210.4 652.7 seedsm
2
yr
1
), and controls were lowest (87.1
613.9 seedsm
2
yr
1
). Greater tree seed deposition in the plantations was due to birds (0.51
60.18 seedsm
2
d
1
), not bats (0.07 60.03 seedsm
2
d
1
). Seed rain was primarily small-
seeded, early-successional species. Large and medium islands received twice as many
zoochorous tree seeds as small islands and areas away from island edges, suggesting there is
a minimum island size necessary to increase seed deposition and that seed rain outside of
planted areas is strongly reduced. Planting design was more important for seed deposition
than amount of forest cover within the surrounding 100- and 500-m radius areas. Establishing
plantations and large islands facilitates the arrival of early-successional tree seeds and
represents a broadly applicable strategy for increasing seed rain on abandoned agricultural
lands. However, more intensive restoration approaches may be necessary for establishment of
dispersal-limited species.
Key words: forest succession; nucleation; plantations; restoration; rural landscapes; seed dispersal; seed
rain; tree islands; tropical montane forest.
INTRODUCTION
Tropical forest restoration is an important component
of global strategies to conserve biodiversity and
sequester atmospheric carbon, particularly in areas
where forests have been extensively cleared and socio-
economic forces are causing land to be removed from
agricultural uses at large scales (Lamb et al. 2005,
Wright and Muller-Landau 2006, Chazdon 2008).
Natural recovery of tropical forests on abandoned and
degraded agricultural lands can be limited by a range of
factors beginning with severely restricted dispersal of
forest seeds into cleared areas (e.g., Duncan and
Chapman 1999, Holl 1999, Wijdeven and Kuzee 2000,
Rodrigues Da Silva and Matos 2006). The lack of seed
dispersal by animals in particular is a major impediment
to forest recovery because soil seed banks are rapidly
depleted by intensive land uses such as grazing, tillage,
and burning (Zimmerman et al. 2000, Cubina and Aide
2001), and animal-dispersed seeds decreases dramatical-
ly within only a few meters of forest edges (e.g., Gorchov
et al. 1993, Holl 1999, Dosch et al. 2007). This limited
dispersal is particularly problematic in wet tropical
forest ecosystems where 60–90%of forest canopy trees
and nearly 100%of shrubs and subcanopy trees are
adapted for animal dispersal (Howe and Smallwood
1982).
Planting native tree seedlings in plantations is the
predominant strategy for accelerating forest succession
(Lamb et al. 2005, Chazdon 2008). Trees attract seed
dispersing animals by providing perching and roosting
sites, habitat for foraging, and cover from predators.
Manuscript received 27 April 2009; revised 23 July 2009;
accepted 10 August 2009. Corresponding Editor: E. Cuevas.
3
E-mail: cole.rebeccaj@gmail.com
1255
Trees also facilitate movement of birds and bats through
agricultural landscapes (McDonnell and Stiles 1983,
Nepstad et al. 1991, Estrada et al. 1993, Guevara and
Laborde 1993, Wunderle 1997, Harvey 2000b). In
addition, trees serve to overcome a range of barriers to
seedling establishment by shading out competitive
grasses (Parrotta 1992, Guariguata et al. 1995), amelio-
rating microclimatic conditions (Guevara et al. 1992,
Parrotta 1995, Holl 1999, Nepstad et al. 1999), and
improving soil chemical and physical properties
(Sanchez et al. 1985, Montagnini and Sancho 1994,
Ashton et al. 1997). Establishing and maintaining
plantations of trees, however, can be an expensive and
labor-intensive endeavor. Accordingly, it is becoming
increasingly important to develop strategies that are
both ecologically and economically effective alternatives
for restoring large areas of abandoned lands with limited
available resources.
Planting trees in patches or ‘‘islands’’ simulates the
nucleation model of succession (sensu Yarranton and
Morrison 1974) and may be a less expensive restoration
approach than establishing plantations. Nucleation
occurs when early-successional vegetation establishes
in patches that spread outward clonally and/or by
facilitating the seed dispersal and establishment of later-
successional species (Yarranton and Morrison 1974 ).
Previous research on abandoned tropical agricultural
lands suggests that this model may apply. For example,
numerous studies have demonstrated that remnant trees
(e.g., Guevara et al. 1992, Duncan and Chapman 1999,
Otero-Arnaiz et al. 1999, Galindo-Gonzalez et al. 2000,
Slocum 2001, Hooper et al. 2005, Schlawin and Zahawi
2008), patches of early-colonizing shrubs (Vieira et al.
1994, Holl 2002, Puyravaud et al. 2003), and rotting logs
(Peterson and Haines 2000, Slocum 2000) enhance seed
rain and seedling establishment in their immediate
vicinities.
The nucleation model shows a great deal of promise
as a restoration tool because it simulates a pattern of
natural recovery. However, only a few studies have
tested its application to restoration (Robinson and
Handel 2000, Zahawi and Augspurger 2006, Benayas
et al. 2008). The only previous study testing the role of
planted tree islands in facilitating recovery of tropical
forest demonstrated that bird activity, seed rain, and
seedling establishment were elevated in tree islands
compared to open pasture within the first two years after
planting (Zahawi and Augspurger 2006).
The potential for island expansion, a critical compo-
nent of the nucleation model (Auld and Coote 1980), is
not well understood. The majority of previous studies
which suggest support for the nucleation model have
demonstrated only higher establishment and survival of
seedlings below shrubs or trees compared to open areas
(Debussche and Isenmann 1994, Verdu´ and Garcı
´a-
Fayos 1996, Carriere et al. 2002, Franks 2003, Garcia
and Obeso 2003, Russell-Smith et al. 2004; but see
Schlawin and Zahawi 2008), rather than recording
whether there is enhanced seed deposition beyond the
edge of the nuclei in the potential expansion zone as
reported by Zahawi and Augspurger (2006). Last, no
evaluations have compared the restoration potential of
small tree nuclei to larger plantations.
Another issue to consider in applying the nucleation
model to restoration is the importance of island size. A
few studies have found higher seed rain and tree seedling
density and diversity in larger rather than smaller
patches (Campbell et al. 1990, Cook et al. 2005,
Zahawi and Augspurger 2006), whereas others have
found no effect (Guevara et al. 1986, Robinson and
Handel 2000, Holl 2002). Larger islands are likely to be
more attractive to dispersers because they are more
highly visible and provide greater amounts of habitat for
cover and foraging. Prior research suggests that birds
are more likely to visit larger islands and stay there
longer, potentially resulting in more seed dispersal and
seedling establishment (Zahawi and Augspurger 2006,
Fink et al. 2009). Nonetheless, concrete information is
lacking on what constitutes a ‘‘sufficiently large’’ island
and whether there is a minimum critical size threshold
for tree islands to function as effective regeneration
nuclei.
The location of a restoration site within the landscape
often plays a critical role in determining seed rain
dynamics. Past studies in second growth habitats in the
tropics have demonstrated the importance of remnant
forest proximity for dispersal and establishment of
zoochorous plants (Thomlinson et al. 1996, Harvey
2000a,b, Zanne and Chapman 2001, Chinea 2002,
Ferguson et al. 2003), whereas others have shown no
such trend (e.g., Guevara et al. 1986, Aide et al. 1996,
Zahawi and Augspurger 2006, Dosch et al. 2007, Pejchar
et al. 2008). If landscape-scale factors, such as the
amount of forest cover in the surrounding landscape, are
important drivers of seed dispersal patterns, seed density
and diversity should be greater in areas with more forest
cover nearby. Conversely, if seed dispersers preferen-
tially utilize restored areas planted with trees, then local-
level factors such as planting design may more strongly
influence the nature of seed dispersal. Understanding the
importance of the extent of surrounding forest on
recovery is critical to prioritizing areas and methods of
restoration in fragmented landscapes.
Key groups of seed dispersers may respond differently
to restoration approachesand the amount of surrounding
forest cover. In areas that have been altered extensively by
human activity, larger vertebrates are often rare or
absent, and small-bodied animals, such as birds and bats,
are the primary seed dispersers (e.g., Estrada et al. 1993,
Nepstad et al. 1996, Galindo-Gonzalez et al. 2000,
Barrantes and Pereira 2002, Martinez-Garza and
Gonzalez-Montagut 2002, Guevara et al. 2004, Griscom
et al. 2007). Previous studies indicate that many forest
birds avoid open areas (DaSilva et al. 1996) and use
woody vegetation for foraging and movement through
agricultural areas (Perfecto et al. 1996, Estrada et al.
R. J. COLE ET AL.1256 Ecological Applications
Vol. 20, No. 5
2000), whereas bats are more likely to frequent open
habitat (Deforesta et al. 1984, Medina et al. 2007).
However, the relative roles of each of these important
disperser groups have rarely been compared, particularly
in restoration settings.
The objective of this study was to test the potential of
applied nucleation as a restoration strategy in a region
characteristic of Central American agricultural land-
scapes. We measured seed rain for 1.5 years beginning
two years after applying three 50 350 m restoration
treatments: control (natural regeneration), islands
(planting native tree nuclei of three sizes: 4 34m,83
8m,12312 m) and plantation (planting uniformly with
tree seedlings). In order to evaluate patterns of seed rain
across a range of conditions and make results general-
izable on a meaningful spatial scale, we replicated the
study at 11 sites across a 100-km
2
area of southern
Costa Rica. Sites were distributed across a gradient of
remnant primary forest cover in the surrounding
landscape, enabling us to measure the effect of
proximity to seed sources on seed rain. The specific
goals of this research were to (1) compare the effect of
experimental treatments on the density and species
composition of seed rain; (2) determine the effect of
island size on the density and species composition of
seed rain; (3) measure the potential for island expansion
by quantifying seed rain at several distances from island
edges; (4) compare the relative seed dispersal contribu-
tions of bats and birds; and (5) evaluate whether local
restoration strategy or amount of surrounding remnant
forest more strongly affected patterns of seed deposi-
tion.
METHODS
Study area.—The study was conducted from February
2006 through August 2008 at 11 ;1-ha experimental
sites distributed across a ;100-km
2
area between the
Las Cruces Biological Station (LCBS; 88470700 N,
8285703200 W) and the town of Agua Buena (884404200
N, 8285605300 W) in southern Costa Rica (Appendix C).
The forest in this region is classified as a tropical
montane rain forest by Holdridge et al. (1971). Sites
range in elevation from 1100 to 1300 m above sea level
and mean annual rainfall is ;3500 mm with a distinct
dry season from December to March. The region was
largely forested until approximately 60 years ago when
government-sponsored immigration led to the develop-
ment of land for small-scale agriculture. Between 1950
and 1980, forest was cleared extensively for coffee
production; however, with the collapse of the coffee
market beginning in the early 1990s, much of the land
under agriculture was converted to cattle pasture
(Rickert 2005). As is typical of much of Central
America, the landscape is a highly fragmented mosaic
of small remnant forests, patches of active agriculture,
fallow plots, and pasture. Estimates show that less than
27%of the land in a 15-km radius surrounding LCBS
remains forested (Daily et al. 2001).
Experimental design.—Seven experimental sites were
established in 2004 and four in 2005 (hence 2.5 and 1.5
years of tree growth prior to initiating this study in
2006); site establishment was spread over two years due
to the logistics of setting up a large-scale project.
Because of high variability in the rates of tree growth,
substantial overlap in mean tree height existed between
the two planting years at the start of seed rain
measurements (105.0– 457.4 cm for 2004 sites and
124.0–221.9 cm for 2005 sites; K. D. Holl and R. A.
Zahawi, unpublished data). The sites are separated by a
minimum of 500 m and are representative of lands being
removed from agriculture in the region. All of the sites
had been farmed for 18 years, usually first for coffee
and then pasture, and most are steeply sloping (see
Appendix B for details). Forest cover within 100 and 500
m radii from the center of each experimental plot was
hand digitized from ortho-rectified 2005 aerial photo-
graphs and comprehensively ground checked. Forest
cover spans a range from ,1%to 66%within a 100-m
radius surrounding the plots and from 9%to 89%in a
500-m radius (Appendix C).
At each site, we established three 0.25-ha (50 350 m)
plots separated by .5 m. Each plot received one of three
treatments: control, island, or plantation (Fig. 1; see
Plate 1). No trees were planted in the control treatment
to allow for natural regeneration. The plantation
treatment was uniformly planted with tree seedlings.
The island treatment was planted with six islands of tree
seedlings (hereafter referred to interchangeably as nuclei
or islands) of three sizes: two each of 4 34m,838m,
and 12 312 m. Nuclei were separated by a minimum of
8 m (Fig. 1). The same planting density of seedlings was
used in both treatments; however, 313 individuals were
planted in plantations and 86 in islands. Two of the
island plots differed slightly from the standard layout
due to space constraints on the available land: the rows
of nuclei were offset at one site and were separated by
more than the normal distance at another site.
Preliminary analyses with and without data from these
two sites showed similar results; therefore, data were
included in the results presented.
Following clearing of all aboveground vegetation in
each plot, we planted seedlings of four tree species that
have high regional survival, rapid growth, and extensive
canopy development within a couple years (Nichols et
al. 2001, Jime
´nez et al. 2002, Carpenter et al. 2004,
Calvo-Alvarado et al. 2007). Two native hardwood
species, Terminalia amazonia (Combretaceae) and
Vochysia guatemalensis (Vochysiaceae), produce valu-
able timber and have high native woody species
establishment in their understory (Cusack and
Montagnini 2004). Two naturalized softwood species,
Erythrina poeppigiana and Inga edulis (both Fabaceae),
are fast-growing N-fixing species widely used in agricul-
tural intercropping systems to provide shade and
increase soil nutrients, and have extensive branching
architecture and fruit (Inga) that attract birds
July 2010 1257SEED RAIN UNDER TREE ISLANDS
(Pennington and Fernandes 1998, Nichols et al. 2001,
Jones et al. 2004). In both island and plantation plots,
seedlings were planted in alternating rows of hardwoods
(Terminalia and Vochysia) and softwoods (Erythrina
and Inga; Fig. 1). Species were planted alternately 4 m
apart, and rows were separated by 2 m and offset by 2 m
so that all seedlings were separated by a minimum of 2.8
m which is standard for reforestation in the region
(Calvo-Alvarado et al. 2007).
All of the plots (including the control) were cleared to
ground level by machete at ;3-month intervals for the
first 2.5 years to allow planted tree seedlings to grow
above existing grasses and forbs. Vegetation at the
termination of clearing was dominated by a mix of
native and nonnative grasses, including Axonopus
scoparius,Paspalum spp., Pennisetum purpureum, and
Urochloa brizantha, and ruderal herbs, such as
Heterocondylus vitalbae,Pteridium arachnoides,and
Spermacoce assurgens (K. D. Holl and R. A. Zahawi,
unpublished data).
Seed trap design.—Seed traps were constructed from
fine-gauge (0.530.5 mm) mosquito netting suspended in
an inverted pyramid from circular wire hoops and
mounted on 50 cm tall legs (trap collecting area ¼0.25
m
2
). One or two large rocks were placed in the bottom of
each trap to prevent the mesh being turned inside out by
wind. Seeds dispersed by ground-dwelling mammals
were excluded from collection; previous research on seed
dispersal by mammals in the study plots showed these
events to be extremely rare (Cole 2009a).
Data collection.—We collected seeds twice monthly in
12 traps in each control and plantation plot and in 22
traps in island plots (Fig. 1). More traps were placed in
island plots to enable comparisons of patterns of seed
FIG. 1. (A) Experimental design, illustrating one possible layout of a site with three 50 350 m plots. Treatment and island
locations were randomized with the constraint in the latter that there was always one island of each size (4 34m,838m,12312
m) on the left and right sides of the plot. Gray indicates areas that were planted with trees separated by 2.8 m along the diagonal
and 4 m in rows. (T ¼Terminalia amazonia,V¼Vochysia guatemalensis,E¼Erythrina poeppigiana,I¼Inga edulis). Black squares
show the locations of seed traps. (B) Placement of seed traps (black squares) in an island plot. Groups of four traps were placed at
2 m and 4 m from the edges of medium and large islands and along midpoints between islands (;8–12 m from island edges).
R. J. COLE ET AL.1258 Ecological Applications
Vol. 20, No. 5
rain within the plot, including island size and expansion.
To compare seed rain in islands of different sizes, one,
three, and six traps were randomly placed within one
small, medium, and large island, respectively (sampling
intensity ;1%). Selection of the island that received
traps was random; however, if a selected island had poor
planted tree survival (,50%) the better-developed of the
two islands was selected. This was done because a
primary objective of the study was to assess the effect of
nuclei size and poor tree survival substantially reduces
island area. To quantify seed rain in the unplanted area
adjacent to each tree island (expansion zone), two traps
each were placed at both 2 and 4 m away from the base
of the outer trees of medium and large islands. Lastly,
four traps were placed along the center of the plot
between rows of islands (;12 m from island edges; Fig.
1). Seed rain near the edge of small islands was not
quantified. Traps in the control and plantation were
placed in groups of three at the edges of four 8 38m
permanent sampling plots, each in a different quadrant
of the plot (Fig. 1). This placement enabled seed traps to
be paired with other research monitoring changes in
vegetation structure and species composition. Fig. 2
shows the differences between island interior and the
expansion zone outside of the planted areas.
To assess the relative contributions to the seed rain by
birds and bats, we placed 12 daytime and 12 nighttime
traps in control and plantation treatments at a subset of
sites. Traps were either opened or closed at ;05:00
06:00 and ;17:00–18:00 each day. Given the time
necessary to open and close traps twice daily within a
narrow time window and the travel distance among
sites, we compared bat and bird dispersal for six weeks
from June to July 2007 (peak of the fruiting season) and
at four sites only. We surveyed two sites at a time in
alternating weeks, so we collected data at each site for
three weeks. Trap contents were collected once weekly.
To examine how differences in tree height affected
seed dispersal, we measured the height of all planted
trees in island plots and approximately one-third of trees
(randomly selected) in the plantations (K. D. Holl and
R. A. Zahawi, unpublished data). Percent overhead cover
was measured near each seed trap in the middle of the
study (June–July 2007) using a spherical densiometer
(mean percent cover 6SE in each site: control ¼6.1 6
1.0; island ¼9.5 62.2; plantation ¼54.5 610.9). In
2007, we also censused vegetation in four 16 316 m
plots at each of six remnant forest patches adjacent to
high forest cover sites in order to establish a reference
baseline for mid–late successional species that could
serve as sources of seeds (R. A. Zahawi and K. D. Holl,
unpublished data).
Seed collection and identification.—The contents of all
traps were collected twice monthly from February 2006
to September 2008 (18 months); day- and nighttime
traps were collected weekly as described above. Traps
that were damaged by wind, livestock, or humans were
excluded from the data for that collection period. Heavy
leaf litter and branches that fell into traps were brushed
to remove any seeds adhering to their surface and
discarded in the field. Remaining trap contents were
placed in manila envelopes and dried at ;658C for
several days. Seeds were then sorted, identified, and
counted using hand lenses (103magnification) and a
dissecting microscope for very small seeds. All seeds
were quantified except for grasses, because the purpose
of this study was to examine factors limiting forest
regeneration rather than species dominant in abandoned
pastures (i.e., introduced pasture grasses). Fruiting
FIG. 2. (Top) A medium island with seed traps placed in the
expansion zone in the grass and (bottom) the interior of a large
island.
July 2010 1259SEED RAIN UNDER TREE ISLANDS
plants were collected throughout the region on an
ongoing basis to establish a reference collection, and
seeds were identified to the lowest possible taxonomic
level. Further identifications were made at the Las
Cruces Biological Station herbarium (LCBS) and by
consultation with several botanists. Vouchers of each
species were stored in 70%alcohol and were deposited at
LCBS.
Data analyses.—To compare differences in seed rain
species composition among sites, we characterized seeds
by (1) growth form (herbs, shrubs, trees, lianas, and
unknowns); (2) successional stage (ruderal herbs, early
successional, and mid–late successional); and (3) dis-
persal mode (animal, wind, gravity/explosive). The
category ‘‘ruderal herbs’’ is composed of weedy species
common throughout active agricultural lands. Although
some of the planted Erythrina poeppigiana and Inga
edulis trees set seed during the course of the study, no
seeds of these species were recorded in traps. Therefore,
the ‘‘tree’’ categories refer to seeds that were dispersed
into plots from outside sources and do not reflect input
from planted species.
To equalize the number of traps across treatments for
plot-level analyses, 12 traps (out of 22) from each island
plot were selected in proportion to planted and
unplanted areas in the treatment for comparison with
control and plantation treatments (four traps in tree
island interiors, two traps at 2 m, and three traps each at
4 and 12 m from nuclei edge; Fig. 1). All island plot
traps were used for analyses of nuclei size and distance
from nuclei edge within island plots. In all analyses of
seed density we used averaged data from the 1.5-year
study for the categories described in the previous
paragraph. Units are either seedsm
2
yr
1
or
seedsm
2
d
1
(bird and bat data only). We also
compared the total number of species (species density)
recorded in each plot for each category.
Our experiment was set up as a randomized complete
block design with site as the blocking factor. Preliminary
analyses indicated no significant differences in seed
densities between sites planted in 2004 and 2005 (P.
0.05 in all cases), so sites from both years were combined
for all analyses. We initially examined the effects of local
restoration treatment (control, island, and plantation)
and landscape-level characteristics (percent forest cover
in the surrounding landscape) for all seed categories
using the following general linear model:
y¼Treatment þBlock þ%Forest Cover
þTreatment 3%Forest Cover þerror:
Percent forest cover includes both primary and
secondary forests (10 years growth), since preliminary
analyses showed no differences in trends when they were
analyzed separately. Separate analyses were performed
using measures of surrounding forest cover at 100 and 500
m radii from the center of each plot, as we had insufficient
replication to include both distances and their interactions
with treatments in the model simultaneously. As forest
cover at both distances did not significantly affect overall
seed or species densities in any of the broad categories
outlined at the beginning of this section (P.0.10), they
were removed from the model. We used the same model
TABLE 1. Percentage of total, total number of species, and mean density (6SE) of seeds by habit, successional stage, and dispersal
mode in each experimental treatment.
Category
Control Island
Percentage
of total
Number
of species
Density
(seedsm
2
yr
1
)
Percentage
of total
Number
of species
Density
(seedsm
2
yr
1
)
Habit
Herb 59 34 952.7
a
6547.5 46 33 710.3
a
6334.1
Shrub (anem.) 5 2 79.4
a
635.2 10 2 156.4
a
675.4
Shrub (zoo.) 13 19 201.4
a
662.7 22 19 345.9
a
6116.4
Tree (anem.) 13 4 202.1
a
6102.4 3 6 45.4
a
617.9
Tree (zoo.) 5 12 87.1
a
613.9 14 16 210.4
ab
652.7
Liana ,1 5 6.0
a
63.6 ,1 7 7.6
a
63.6
Unknown 4 34 62.6 620.7 5 33 76.7 640.4
Successional stage
Ruderal herb 67 38 1075.1
a
6556.0 55 35 851.2
a
6390.0
Early 26 30 420.3
a
66.8 36 34 556.2
a
6138.2
Mid–late (zoo.) ,1 6 3.7
a
62.0 ,1 10 11.5
a
67.1
Mid–late (anem.) 1 3 10.7
a
69.0 ,1 5 1.4
a
60.7
Dispersal mode
Animal 24 80 378.9
a
677.9 44 90 679.0
a
6151.4
Wind 71 26 1135.0
a
6544.6 49 23 761.7
a
6393.7
Gravity 5 6 87.5
a
628.8 7 6 112.0
a
644.8
All seeds 40 112 1601.4
a
6609.1 39 118 1552.7
a
6465.3
Notes: Values for the control and plantation are for all 12 traps in each treatment combined. Values for the island plot are from
12 traps (four within planted areas, eight in unplanted areas). Anemochorous (anem.) and zoochorous (zoo.) tree, shrub, and mid–
late-successional seeds are shown separately. Means with the same superscript letter are not significantly different (P,0.05) using
Tukey’s hsd.
R. J. COLE ET AL.1260 Ecological Applications
Vol. 20, No. 5
to test for local and landscape-level effects on the seed
densities of the seven most common zoochorous and
anemochorous tree species (mean .10 seedsm
2
yr
1
;
Table 2). Because tree growth varied greatly among sites,
we tested whether tree cover influenced the arrival of
zoochorous tree seeds by regressing seed densities in each
island and plantation treatment against tree height and
overhead cover. We did not do this analysis for controls
because there were no trees planted in these plots, and
there were very few naturally established trees .2 m tall
during the period of this study.
The effects of nuclei size and distance from nuclei edge
on seed rain densities within the island treatment were
compared using a randomized block one-way analysis of
variance (ANOVA) with location within island plot
(small, medium, or large island; 2 m, 4 m, or 12 m from
edge) as the explanatory variable. We focused this
analysis on zoochorous trees and shrubs as they are of
key importance to nuclei expansion and patterns of
anemochorous seed rain should not be affected by
plantings. Finally, we compared the relative contribu-
tions to the seed rain of birds and bats using a
randomized-block two-way ANOVA with treatment
(control or plantation), time of day (day or night), and
their interactions.
In all analyses, seed density data were square-root
transformed and the residuals examined to determine
that the data met assumptions of normality and
homogeneity of variances (Zar 1996). Species density
and percent forest cover data were normally distributed
and required no transformations. We used Tukey hsd
post hoc tests to determine significant differences among
treatments when appropriate. Throughout, means 61
SE are reported and P,0.05 is considered significant.
All analyses were conducted using Systat 12.0 (Systat
Software 2007).
RESULTS
Seed rain characteristics.—A total of 251 768 seeds
were collected in control, island, and plantation
treatments over the 1.5-year sampling period. Seeds
represented 168 species from at least 43 families. Nearly
half (45.7%) were Asteraceae, most of which are
anemochorous. The next most common families were
Solanaceae (10.6%), Melastomataceae (7.9%), and
Urticaceae (7.2%), all of which are zoochorous. Of all
seeds collected, 77 species (comprising ,6%of the total
number of seeds) could not be identified; it was possible,
however, to determine dispersal mode in most cases.
Herbs, mainly common ruderal species, accounted for
44%of all seeds and 31%of all species. Forty-eight
percent of all seeds and 75%of all species were
zoochorous, whereas 44%of seeds and 20%of all
species were anemochorous. Fourteen percent of seeds
were zoochorous trees (22 species) and 5%were
anemochorous (6 species). A small fraction of the seeds
that arrived in the plots (0.4%) were classified as mid-
late successional species (Table 1). Appendix A provides
a summary of all species of seeds grouped by growth
form, successional stage, and dispersal mode.
The most commonly dispersed zoochorous and
anemochorous trees were small-seeded (maximum di-
ameter ,2.0 mm) early-successional species commonly
found in edge habitat, along roadsides, and fallow fields.
Of the 146 tree species surveyed in forests adjacent to six
of the experimental sites (R. A. Zahawi and K. D. Holl,
unpublished data), only 11 species were recorded in the
seed traps. Five early-successional zoochorous tree
species accounted for 99.3%of all animal-dispersed tree
seeds collected (Table 2). Similarly, a single anemocho-
rous tree species, Heliocarpus appendiculatus, accounted
for 89.1%of wind-dispersed tree seeds (Table 2).
As with previous studies, we noted distinct seasonal
patterns in the seed rain. The vast majority of anemocho-
rous tree seeds (88.5%) fell during the dry season between
December and March and 78.4%of zoochorous tree
species were deposited during the early part of the wet
season from April to July (Fig. 3). The total number of
seeds dispersed during the dry season and the first half of
the wet season in the two years was similar.
Treatment and landscape effects.—The amount of
forest cover within 100 and 500 m radii of the
experimental plots did not significantly affect either the
seed rain or species density of any category (F3.2, df
¼2, 17, P.0.05 in all cases) and was therefore removed
from the statistical model for these analyses. Plantations
received nearly three times as many zoochorous tree
seeds (266.1 664.5) as controls (87.1 61 3.9) (F¼3.8,
df ¼2, 20 P¼0.0405); island plots received an
intermediate number (210.4 652.7) but were not
significantly different from either treatment due to high
variances among sites (Table 1). Species densities did not
TABLE 1. Extended.
Plantation
Percentage
of total
Number
of species
Density
(seedsm
2
yr
1
)
25 32 208.8
a
633.4
2 2 20.4
a
69.3
24 20 197.8
a
641.2
9 6 71.5
a
631.6
32 15 266.1
b
664.5
,1 5 3.7
a
61.0
6 26 50.3 618.2
30 37 243.3
a
637.1
51 27 417.3
a
654.5
,1 10 1.9
a
60.7
,1 5 1.7
a
60.8
67 79 551.3
a
696.3
31 21 252.6
a
646.5
2 5 15.8
a
64.8
21 106 551.3
a
696.3
July 2010 1261SEED RAIN UNDER TREE ISLANDS
significantly differ among treatments for any other seed
category.
Seed rain density of Cecropia peltata, a commonly
dispersed zoochorous tree, was higher in plantation plots
(but not other treatments) with more forest cover in the
surrounding 100 m area (forest cover 3treatment
interaction; F¼3.6; df ¼2, 17; P¼0.0486). Deposition
of this species in plantations was five and two times
greater than in control and island treatments, respectively
(Table 2). Seed density of the other common zoochorous
trees was not related to forest cover (F3.4, df ¼2, 17, P
.0.05). Cecropia obtusifolia was dispersed in greater
amounts to the plantation compared to the control plot
(F¼3.8; df ¼2, 20; P¼0.0397), and Conostegia
rufenscens was dispersed in greater densities to both of
the planted treatments compared to the controls (F¼6.4,
df ¼2, 20, P¼0.0071; Table 2). Seed rain densities of
individual anemochorous tree species did not differ as a
function of treatment or percent forest cover.
As is typical for seed rain data, there was high among-
site variability in patterns of seed deposition. For
example, nearly 4.5 times as many herbaceous seeds
were deposited in control plots compared to plantations,
but 59.9%of all herbaceous seeds collected in control
plots arrived at a single site. Similarly, there was a
tendency for more anemochorous tree seeds to arrive in
control plots compared to other treatments but this was
likely due to very local-level factors. Specifically, the
presence of several Heliocarpus appendiculatus trees near
the control plot of one site accounted for 31.6%of
anemochorous tree seeds collected. Surprisingly, neither
tree height nor percent overhead cover explained a
significant percentage of among site variance in levels of
FIG. 3. Seasonal patterns of zoochorous and anemochorous tree seeds falling in the experimental sites.
TABLE 2. Mean seed density (6SE) of the seven most commonly dispersed tree species in control, island, and plantation
treatments (includes all species with mean density .3 seedsm
2
yr
1
).
Species Family Dispersal mode
Density (seedsm
2
yr
1
)
Pasture Islands Plantation
Heliocarpus appendiculatus Tiliaceae wind 191.4
a
6100.1 44.12
a
618.0 69.8
a
631.8
Cecropia peltataUrticaceae bird, bat 21.3
a
68.0 43.8
a
69.7 105.2
b
646.0
Cecropia obtusifolia Urticaceae bird, bat 17.7
a
66.3 51.0
b
616.1 44.3
b
612.4
Conostegia rufescens Melastomataceae bird, bat 15.8
a
68.9 18.1
ab
67.8 70.1
b
634.7
Miconia trinervia Melastomataceae bird, bat 17.0
a
69.2 43.3
a
619.0 22.3
a
67.8
Conostegia xalapensis Melastomataceae bird, bat 11.0
a
62.8 45.7
a
622.7 22.0
a
64.4
Ulmus Mexicana Ulmaceae wind 9.9
a
68.9 0.8
a
60.6 0.5
a
60.2
Notes: Seed density is ranked by decreasing overall abundance. All of the zoochorous species and one of the anemochorous
species (Heliocarpus appendiculatus) are common early-successional species. Ulmus mexicana is a late-successional species often
occurring as a remnant in agricultural lands. All trees are small-seeded (,2.0 mm diameter). Treatments with the same superscript
letter are not significantly different (P,0.05) based on Tukey’s hsd multiple comparisons.
Seed density of Cecropia peltata in plantations, but not other treatments, increased with increasing forest cover within a 100-m
radius.
R. J. COLE ET AL.1262 Ecological Applications
Vol. 20, No. 5
zoochorous seed density in plantation or island plots (r
2
,0.03; P.0.05 in all cases).
Island size and distance from island edge.—More
zoochorous tree seeds were dispersed to large and
medium-sized islands than to either small islands, or
areas in the expansion zone at 2, 4, or 12 m away from
island edges (F¼11.2, df ¼5, 50 P,0.0001; Fig. 4).
There was no significant difference between the zoo-
chorous tree seed density in small islands and traps in
the expansion zone or among traps at different distances
from island edges. Similarly, more zoochorous tree
species arrived in the large island compared to the small
island and all the expansion zone locations (F¼9.0; df ¼
5, 50; P,0.0001); medium islands received more species
than small islands but were similar to all other locations
within the island plots.
Dispersal of zoochorous shrubs showed a similar but
less consistent pattern (Fig. 4). Both large and medium
islands received more shrub seeds than several of the
expansion zone locations (F¼4.1; df ¼5, 50; P¼0.0032)
but were similar to the small islands. More zoochorous
shrub species fell in large compared to small islands and
all other expansion zone locations, whereas medium
islands received more shrub species than small islands
but were similar to all other locations within the island
plots (F¼11.5; df ¼5, 50; P,0.0001; Fig. 4).
FIG. 4. (A) Zoochorous tree and shrub seed and (B) species densities at each location within island plots. Treatments with the
same letter are not significantly different (P,0.05) based on Tukey’s hsd multiple comparisons among trees (lowercase letters) and
shrubs (uppercase letters). Error bars are þSE.
July 2010 1263SEED RAIN UNDER TREE ISLANDS
Seed dispersal by birds and bats.—Seed dispersal
during the day (birds) and at night (bats) differed
significantly as a function of planting treatment (treat-
ment 3time of day interaction; F¼8.2; df ¼1, 9; P¼
0.0186; Fig. 5). Deposition of tree seeds by bats was
similar in controls and plantations, whereas birds
dispersed nearly seven times more seeds in plantations
than control plots (islands were not assessed in this
study). As a result of the greater seed rain from birds,
more zoochorous tree seeds were deposited in the
plantations than the control plots in the four sites
included in this portion of the study (F¼13.1; df ¼1, 9;
P¼0.0056). Zoochorous shrub seeds were also
dispersed in greater quantities to plantations than
controls (F¼15.7; df ¼1, 9; P,0.0033) but bats and
birds contributed in similar quantities to the shrub seed
rain (Fig. 5).
DISCUSSION
General overview.—The results of this study lead to
several conclusions regarding applied nucleation as a
restoration strategy and patterns of forest succession on
post-agricultural lands. First, we found that planting
trees both in plantations and islands substantially
enhanced the arrival of small tree seeds dispersed by
birds within the first 2–4 years after planting, but did not
affect arrival by wind, gravity, or bats. Second, our
results concur with the findings of previous studies
showing that the dispersal of primary forest species is
strongly limited in fragmented landscapes regardless of
the restoration strategy used (Holl 1999, Harvey 2000b,
Dosch et al. 2007). Third, we found that the amount of
forest cover within the surrounding 100- and 500-m
landscape had a relatively minimal effect on seed
dispersal compared to tree planting strategy during this
early stage in succession. Finally, there appears to be a
minimum critical size threshold for tree islands neces-
sary to increase the deposition of zoochorous tree seeds.
Restoration treatments.—Based on past studies, we
predicted that dispersal of zoochorous seeds would be
higher in plantation and island treatments than in the
control plots because the areas planted with trees would
be attractive to frugivores and would provide more
habitat for foraging and protection (e.g., Parrotta et al.
1997, Lamb et al. 2005, Zahawi and Augspurger 2006,
Orozoco Zamora and Montagnini 2007, Fink et al.
2009). Indeed, three times as many zoochorous tree
seeds arrived in plantations as in controls. Island plots
were more similar to plantations in terms of the number
of seeds deposited. The number of seeds arriving in the
island plots was probably slightly lower than plantations
at this stage in the development of the nuclei because of
the relatively high proportion of open (;85%)to
planted areas (;15%). In fact, the mean number of tree
seeds per trap collected inside the planted areas of the
island treatments (295.1 690.9) was similar to the
plantations (266.1 664.5), but as the results of the
within-plot study of the island treatment suggest, seed
deposition beyond the edge of the areas planted with
trees dropped sharply.
Our results also suggest that there is a minimum size
threshold for islands necessary to be attractive to seed
dispersers and to enhance seed rain. Deposition of
zoochorous tree seeds was higher in both the large and
medium islands compared to the small island or any of
the areas in the expansion zone. Although patterns of
shrub seed dispersal were not as sharply defined, large
and medium islands tended to have greater seed rain
compared to areas in the expansion zone. Observations
of bird activity at six of our sites also showed higher bird
visitation and foraging in the large and medium islands
compared to the smaller islands (Fink et al. 2009).
Similarly, Zahawi and Augspurger (2006) found higher
levels of zoochorous tree seed deposition in larger tree
islands (64 m
2
) planted with Gliricidia sepium compared
to smaller islands (16 and 4 m
2
). From a restoration
perspective, these findings suggest that nuclei .64 m
2
are more effective for facilitating forest recovery than
smaller nuclei, at least in the seed-dispersal stage.
FIG. 5. Mean seed density of tree and shrub seed by
disperser and treatment. Treatments with the same letter are not
significantly different (P,0.05) based on ANOVA. Error bars
show þSE. Birds dispersed significantly more tree seeds than did
bats in the plantation treatment (shown by a dagger).
R. J. COLE ET AL.1264 Ecological Applications
Vol. 20, No. 5
We did not find increased levels of tree or shrub seed
deposition in the expansion zone near the outside of
island edges, which is consistent with previous studies
reporting a dramatic drop in the density and diversity of
zoochorous seed rain only a few meters away from the
edge of forest (Holl 1999, Cubina and Aide 2001, Dosch
et al. 2007). Zahawi and Augspurger (2006) found
similar levels of seed density in the interior of planted
tree patches and in a 1-m expansion zone; a smaller area
than we measured. Although studies in distinct habitats
often show that seed rain is higher under remnant trees
and other types of vegetation that provide perching
structures (e.g., Guevara et al. 1986, Harvey 2000b),
there has been relatively little testing to determine
whether seed rain is also enhanced beyond the perimeter
of the vegetation. If increased levels of seed rain occur
only directly beneath established woody vegetation as
this study suggests, then island expansion will occur
when seeds falling near the periphery establish and grow
outwards. It is also possible that patterns of nuclei
expansion will be apparent only over longer periods of
time than measured in this study. For example, Schlawin
and Zahawi (2008) found that remnant trees had a
higher density of tree saplings beneath and adjacent to
their canopies than farther away .20 years after site
abandonment.
Seed dispersal by birds and bats.—The difference in
amount of zoochorous tree seeds entering the plantations
vs. control plots appears largely to have been due to
enhanced bird activity. Birds dispersed .80%more tree
seeds in plantations than did bats, whereas contributions
to tree seed rain in control plots by both groups of
dispersers was similar. Interestingly, the planting treat-
ment appears to enhance dispersal of shrub seeds by both
birds and bats: shrub seed density was twice as high in
plantations as in controls over the six weeks of this
portion of the study. The importance of these major
disperser groups appears to vary among neotropical
agricultural landscapes. Some studies show that bats
disperse more (Medellin and Gaona 1999), similar
(Galindo-Gonzalez et al. 2000), or fewer numbers
(Harvey 2000b, Gonzales et al. 2009) of species and seeds.
It is possible that bats are more effective dispersers of
shrub than tree seeds in this region (e.g., Muscarella and
Fleming 2007). For example, bats have been shown to be
key dispersers of fast-growing early- and mid-successional
shrubs in Costa Rican lowland forest (Kelm and von
Helversen 2007), dry forests in Panama (Griscom et al.
2007), and wet forest in Mexico (Medellin and Gaona
1999). Further research on the roles of each disperser
group in this region would be informative for develop-
ment of restoration approaches that facilitate seed
dispersal by frugivores.
Landscape-level effects.—A major objective of this
research was to explore the relative roles of local vs.
landscape-level processes driving patterns of seed
dispersal. The local factors of tree planting design
appeared to have a much stronger influence on the
dispersal of zoochorous tree and mid-late successional
seeds than the amount of forest within a 100- or 500-m
radius. This result is not surprising given that tree seed
rain was dominated by only five early-successional
PLATE 1. Photo of one experimental site taken in 2006 at the initiation of the study. The plantation (P) is outlined on the left,
the control (C) is in the center, and the island treatment (I) is on the right. Photo credit: R. J. Cole.
July 2010 1265SEED RAIN UNDER TREE ISLANDS
species (99.5%of total zoochorous tree seeds), none of
which are restricted to forest habitats. Studies examining
seed dispersal and seedling recruitment at distances from
forest greater than 10–25 m have often found no
consistent trends (e.g., Guevara et al. 1986, Slocum
and Horvitz 2000, Slocum 2001, Zahawi and
Augspurger 2006). Likewise, research in the region near
LCBS has shown no correlations between abundance or
diversity of seed rain and forest fragment size (Dosch et
al. 2007), total forest area within 10- to 1000-m buffer
areas, or with distance to the nearest forest fragment
(Pejchar et al. 2008). The structure of vegetation outside
of forests can be important to patterns of bird
movement, and tropical forest birds have often been
shown to frequent the agricultural matrix (e.g., Daily et
al. 2001, Sodhi et al. 2005, Peh et al. 2006, Sekercioglu et
al. 2007). Nearly half of the bird species in our region
have been found in agricultural areas (Hughes et al.
2002), suggesting that the frugivorous species that utilize
this landscape are dispersing seeds similarly to both low
and high forest cover areas.
An apparent exception to this trend was Cecropia
peltata which was positively correlated with increasing
forest cover surrounding the plantations but not the
other treatments. However, it is quite likely that local
reproductive trees influenced these patterns. For exam-
ple, adult Cecropia peltata trees were observed growing
along forest edges and in fallow fields near three
plantation plots where very high levels of seed rain
(135 to 522 seedsm
2
yr
1
) were recorded. Similarly,
several Heliocarpus appendiculatus trees adjacent to a
control plot were in all likelihood the sources of 32.8%
of all seeds collected for this species, and a large remnant
tree near one of the control plots accounted for 79.9%of
all Ulmus mexicana seeds. These observations point to
the importance of local seed sources, as suggested by
other studies showing that most seeds are dispersed
short distances (Duncan and Chapman 1999, Holl 1999,
Mesquita et al. 2001, Ingle 2003, Dosch et al. 2007, del
Castillo and Rios 2008).
We found that the vast majority of animal-dispersed
seeds arriving at our sites were small-seeded, early-
successional species (98.2%). There was little overlap
between seed rain and the species surveyed in the forest
fragments, even though some of these forests were
immediately adjacent to some of our plots. These results
support the findings of previous studies in distinct
tropical ecosystems showing that many primary forest
species and large-seeded species are often absent from
the seed rain in human-altered tropical lands (e.g., Holl
1999, Ingle 2003, Martinez-Garza and Howe 2003,
Dosch et al. 2007, del Castillo and Rios 2008). It may
be necessary to introduce dispersal-limited species
through direct seeding or enrichment planting as a
secondary phase in the restoration effort (Cole 2009b). It
is possible however, that seed rain composition will
change in later stages of forest development in the
treatments. For example, del Castillo and Perez-Rios
(2008) observed that the species richness and abundance
of late-successional seeds increased with the age of
successional stands in Mexican montane forests, due, in
part, to seed production by local plants. As successional
species establish and grow, the plots will become more
structurally complex and diverse, potentially attracting a
broader range of vertebrate dispersers. The extent of
active outplanting vs. allowing succession to proceed
over time will ultimately depend on the objectives of the
restoration effort.
Implications for tropical forest restoration.—Our
results have potentially broad application for restoring
forest in fragmented landscapes in the neotropics. They
show that ‘‘applied nucleation’’ is a promising restora-
tion strategy when sufficiently large nuclei are planted.
In our study, the number of seeds dispersed in island
plots was only slightly lower than in plantations, yet the
cost of planting was a third of that for plantations.
Additionally, we cleared throughout island plots during
the 2.5-year maintenance phase of treatment setup but in
the actual application of the restoration strategy a land
manager could use other maintenance approaches such
as clearing only in the islands, immediately around
seedlings, or along rows of trees which would further
reduce costs and likely increase the rate of island
expansion. Moreover, our results suggest that planting
tree seedlings in both plantations and larger islands, has
broad applicability throughout the landscape given the
lesser importance of proximity to remaining forest
patches—at least during the early stages of succession
and forest recovery.
Ongoing monitoring of naturally establishing seed-
lings in each of the treatments will be important to
determine whether the different levels of zoochorous tree
seed inputs actually result in enhanced recruitment of
woody vegetation. Nuclei expansion in the island plots, a
critical component of the nucleation model of succession,
and the effects of each planting treatment on subsequent
successional trajectories, bear further observation and
should provide key insights into patterns of forest
recovery over time. Attention should also be given to
more intensive restoration approaches, such as enrich-
ment planting or direct seeding of larger-seeded and
later-successional species that fail to recruit because of
strong limitations on seed dispersal beyond forest edges.
ACKNOWLEDGMENTS
We thank R. Gomez for his assistance in collecting and
identifying seeds and F. Oviedo for assistance with species
identification. We are grateful for research help from J. L. Reid,
F. Obando, J. A. Rosales, G. Sady, R. Sniatowski, T. Kehoe,
and other field assistants. D. Morales provided initial GIS
information. C. Augspurger and two anonymous reviewers
provided helpful comments on earlier drafts of the manuscript.
Support for this project was provided by an NSF grant (DEB
0515577), the Marilyn C. Davis Memorial Foundation, the
Earthwatch Foundation, and the UCSC Environmental Studies
Department, as well as a UCSC writing fellowship for the
senior author. We are grateful to the many landowners who
permitted us to conduct this study on their land.
R. J. COLE ET AL.1266 Ecological Applications
Vol. 20, No. 5
LITERATURE CITED
Aide, T. M., J. K. Zimmerman, M. Rosario, and H. Marcano.
1996. Forest recovery in abandoned cattle pastures along an
elevational gradient in northeastern Puerto Rico. Biotropica
28:537–548.
Ashton, P. M. S., S. Gamage, I. Gunatilleke, and C. V. S.
Gunatilleke. 1997. Restoration of a Sri Lankan rainforest:
using Caribbean pine Pinus caribaea as a nurse for
establishing late-successional tree species. Journal of Applied
Ecology 34:915–925.
Auld, B. A., and B. G. Coote. 1980. A model of a spreading
plant-population. Oikos 34:287–292.
Barrantes, G., and A. Pereira. 2002. Seed dissemination by
frugivorous birds from forest fragments to adjacent pastures
on the western slope of Volcan Barva, Costa Rica. Revista
De Biologia Tropical 50:569–575.
Benayas, J. M. R., J. M. Bullock, and A. C. Newton. 2008.
Creating woodland islets to reconcile ecological restoration,
conservation, and agricultural land use. Frontiers in Ecology
and the Environment 6:329–336.
Calvo-Alvarado, J. C., D. Arias, and D. D. Richter. 2007. Early
growth performance of native and introduced fast growing
tree species in wet to sub-humid climates of the Southern
region of Costa Rica. Forest Ecology and Management 242:
227–235.
Campbell, B. M., T. Lynam, and J. C. Hatton. 1990. Small-
scale patterning in the recruitment of forest species during
succession in tropical dry forest, Mozambique. Vegetatio 87:
51–57.
Carpenter, F. L., J. D. Nichols, R. T. Pratt, and K. C. Young.
2004. Methods of facilitating reforestation of tropical
degraded land with the native timber tree, Terminalia
amazonia. Forest Ecology and Management 202:281–291.
Carriere, S. M., M. Andre, P. Letourmy, I. Olivier, and D. B.
McKey. 2002. Seed rain beneath remnant trees in a slash-
and-burn agricultural system in southern Cameroon. Journal
of Tropical Ecology 18:353–374.
Chazdon, R. L. 2008. Beyond deforestation: restoring forests
and ecosystem services on degraded lands. Science 320:1458–
1460.
Chinea, J. D. 2002. Tropical forest succession on abandoned
farms in the Humacao Municipality of eastern Puerto Rico.
Forest Ecology and Management 167:195–207.
Cole, R. J. 2009a. Post-dispersal seed fate of tropical montane
trees in an agricultural landscape, southern Costa Rica.
Biotropica 41:319–327.
Cole, R. J. 2009b. Ecological and socioeconomic aspects of
restoring forest in a tropical agricultural landscape, southern
Costa Rica. Dissertation. University of California, Santa
Cruz, California, USA.
Cook, W. M., J. Yao, B. L. Foster, R. D. Holt, and L. B.
Patrick. 2005. Secondary succession in an experimentally
fragmented landscape: community patterns across space and
time. Ecology 86:1267–1279.
Cubina, A., and T. M. Aide. 2001. The effect of distance from
forest edge on seed rain and soil seed bank in a tropical
pasture. Biotropica 33:260–267.
Cusack, D., and F. Montagnini. 2004. The role of native species
plantations in recovery of understory woody diversity in
degraded pasturelands of Costa Rica. Forest Ecology and
Management 188:1–15.
Daily, G. C., P. R. Ehrlich, and G. A. Sanchez-Azofeifa. 2001.
Countryside biogeography: use of human-dominated habi-
tats by the avifauna of southern Costa Rica. Ecological
Applications 11:1–13.
DaSilva, J. M. C., C. Uhl, and G. Murray. 1996. Plant
succession, landscape management, and the ecology of
frugivorous birds in abandoned Amazonian pastures. Con-
servation Biology 10:491–503.
Debussche, M., and P. Isenmann. 1994. Bird-dispersed seed
rain and seedling establishment in patchy Mediterranean
vegetation. Oikos 69:414– 426.
de Foresta, H., P. Charles-Dominique, C. Erard, and M.-F.
Pre
´vost. 1984. The importance of zoochory during the 1st
stages of forest regeneration after clear-cutting in French-
Guyana. Revue d’Ecologie—La Terre et la Vie 39:369.
del Castillo, R. F., and M. A. P. Rios. 2008. Changes in seed
rain during secondary succession in a tropical montane cloud
forest region in Oaxaca, Mexico. Journal of Tropical Ecology
24:433–444.
Dosch, J. J., C. J. Peterson, and B. L. Haines. 2007. Seed rain
during initial colonization of abandoned pastures in the
premontane wet forest zone of southern Costa Rica. Journal
of Tropical Ecology 23:151–159.
Duncan, R. S., and C. A. Chapman. 1999. Seed dispersal and
potential forest succession in abandoned agriculture in
tropical Africa. Ecological Applications 9:998–1008.
Estrada, A., P. Cammarano, and R. Coates-Estrada. 2000. Bird
species richness in vegetation fences and in strips of residual
rain forest vegetation at Los Tuxtlas, Mexico. Biodiversity
and Conservation 9:1399–1416.
Estrada, A., R. Coatesestrada, D. Meritt, S. Montiel, and D.
Curiel. 1993. Patterns of frugivore species richness and
abundance in forest islands and in agricultural habitats at
Los-Tuxtlas, Mexico. Vegetatio 108:245–257.
Ferguson, B. G., J. Vandermeer, H. Morales, and D. M.
Griffith. 2003. Post-agricultural succession in El Peten,
Guatemala. Conservation Biology 17:818–828.
Fink, R. D., C. A. Lindell, E. B. Morrison, R. A. Zahawi, and
K. D. Holl. 2009. Patch size and tree species influence the
number and duration of bird visits in forest restoration plots
in southern Costa Rica. Restoration Ecology 17:479– 486.
Franks, S. J. 2003. Facilitation in multiple life-history stages:
evidence for nucleated succession in coastal dunes. Plant
Ecology 168:1–11.
Galindo-Gonzalez, J., S. Guevara, and V. J. Sosa. 2000.
Bat- and bird-generated seed rains at isolated trees in pastures
in a tropical rainforest. Conservation Biology 14:1693–1703.
Garcia, D., and J. R. Obeso. 2003. Facilitation by herbivore-
mediated nurse plants in a threatened tree, Taxus baccata:
local effects and landscape level consistency. Ecography 26:
739–750.
Gonzales, R. S., N. R. Ingle, D. A. Lagunzad, and T.
Nakashizuka. 2009. Seed dispersal by birds and bats in
lowland Philippine forest successional area. Biotropica 41:
452–458.
Gorchov, D. L., F. Cornejo, C. Ascorra, and M. Jaramillo.
1993. The role of seed dispersal in the natural regeneration of
rain-forest after strip-cutting in the Peruvian Amazon.
Vegetatio 108:339–349.
Griscom, H. P., E. K. V. Kalko, and M. S. Ashton. 2007.
Frugivory by small vertebrates within a deforested, dry
tropical region of Central America. Biotropica 39:278–282.
Guariguata, M. R., R. Rheingans, and F. Montagnini. 1995.
Early woody invasion under tree plantations in Costa Rica:
Implications for forest restoration. Restoration Ecology 3:
252–260.
Guevara, S., and J. Laborde. 1993. Monitoring seed dispersal at
isolated standing trees in tropical pastures: consequences for
local species availability. Vegetatio 108:319–338.
Guevara, S., J. Laborde, and G. Sanchez-Rios. 2004. Rain
forest regeneration beneath the canopy of fig trees isolated in
pastures of Los Tuxtlas, Mexico. Biotropica 36:99–108.
Guevara, S., J. Meave, P. Morenocasasola, and J. Laborde.
1992. Floristic composition and structure of vegetation under
isolated trees in neotropical pastures. Journal of Vegetation
Science 3:655–664.
Guevara, S., S. E. Purata, and E. Vandermaarel. 1986. The role
of remnant forest trees in tropical secondary succession.
Vegetatio 66:77–84.
July 2010 1267SEED RAIN UNDER TREE ISLANDS
Harvey, C. A. 2000a. Colonization of agricultural windbreaks
by forest trees: effects of connectivity and remnant trees.
Ecological Applications 10:1762–1773.
Harvey, C. A. 2000b. Windbreaks enhance seed dispersal into
agricultural landscapes in Monteverde, Costa Rica. Ecolog-
ical Applications 10:155–173.
Holdridge, L. R., W. C. Grenke, W. H. Hatheway, T. Liany,
and J. A. Tosi, Jr. 1971. Forest environments in tropical life
zones. Pergamon Press, New York, New York, USA.
Holl, K. D. 1999. Factors limiting tropical rain forest
regeneration in abandoned pasture: seed rain, seed germina-
tion, microclimate, and soil. Biotropica 31:229–242.
Holl, K. D. 2002. Effect of shrubs on tree seedling establish-
ment in an abandoned tropical pasture. Journal of Ecology
90:179–187.
Hooper, E., P. Legendre, and R. Condit. 2005. Barriers to
forest regeneration of deforested and abandoned land in
Panama. Journal of Applied Ecology 42:1165–1174.
Howe, H. F., and J. Smallwood. 1982. Ecology of seed
dispersal. Annual Review of Ecology and Systematics 13:
201–228.
Hughes, J. B., G. C. Daily, and P. R. Ehrlich. 2002.
Conservation of tropical forest birds in countryside habitats.
Ecology Letters 5:121–129.
Ingle, N. R. 2003. Seed dispersal by wind, birds, and bats
between Philippine montane rainforest and successional
vegetation. Oecologia 134:251–261.
Jime
´nez, M., F. Rojas, V. Rojas, and L. Rodrı
´guez. 2002.
Arboles maderables de Costa Rica. Instituto Nacional de
Biodiversidad, Heredia, Costa Rica.
Jones, E. R., M. H. Wishnie, J. Deago, A. Sautu, and A.
Cerezo. 2004. Facilitating natural regeneration in Saccharum
spontaneum (L.) grasslands within the Panama Canal
Watershed: effects of tree species and tree structure on
vegetation recruitment patterns. Forest Ecology and Man-
agement 191:171–183.
Kelm, D. H., and O. von Helversen. 2007. Artificial bat roosts
and seed dispersal: A method to enhance forest restoration.
Bat Research News 48:242.
Lamb, D., P. D. Erskine, and J. A. Parrotta. 2005. Restoration
of degraded tropical forest landscapes. Science 310:1628–
1632.
Martinez-Garza, C., and R. Gonzalez-Montagut. 2002. Seed
rain of fleshy-fruited species in tropical pastures in Los
Tuxtlas, Mexico. Journal of Tropical Ecology 18:457–462.
Martinez-Garza, C., and H. F. Howe. 2003. Restoring tropical
diversity: beating the time tax on species loss. Journal of
Applied Ecology 40:423–429.
McDonnell, M. J., and E. W. Stiles. 1983. The structural
complexity of old-field vegetation and the recruitment of
bird-dispersed plant species. Oecologia 56:109–116.
Medellin, R. A., and O. Gaona. 1999. Seed dispersal by bats
and birds in forest and disturbed habitats of Chiapas,
Mexico. Biotropica 31:478–485.
Medina, A., C. A. Harvey, D. S. Merlo, S. Vilchez, and B.
Hernandez. 2007. Bat diversity and movement in an
agricultural landscape in Matiguas, Nicaragua. Biotropica
39:120–128.
Mesquita, R. C. G., K. Ickes, G. Ganade, and G. B.
Williamson. 2001. Alternative successional pathways in the
Amazon Basin. Journal of Ecology 89:528–537.
Montagnini, F., and F. Sancho. 1994. Aboveground biomass
and nutrients in young plantations of indigenous trees on
infertile soils in Costa Rica: implications for site nutrient
conservation. Journal of Sustainable Forestry 4:115–139.
Muscarella, R., and T. H. Fleming. 2007. The role of
frugivorous bats in tropical forest succession. Biological
Reviews 82:573–590.
Nepstad, D. C., C. Uhl, C. A. Pereira, and J. M. C. daSilva.
1996. A comparative study of tree establishment in aban-
doned pasture and mature forest of eastern Amazonia. Oikos
76:25–39.
Nepstad, D. C., C. Uhl, and E. A. S. Serrao. 1991.
Recuperation of a degraded Amazonian landscape: forest
recovery and agricultural restoration. Ambio 20:248–255.
Nepstad, D. C., A. Verissimo, A. Alencar, C. Nobre, E. Lima,
P. Lefebvre, P. Schlesinger, C. Potter, P. Moutinho, E.
Mendoza, M. Cochrane, and V. Brooks. 1999. Large-scale
impoverishment of Amazonian forests by logging and fire.
Nature 398:505–508.
Nichols, J. D., M. E. Rosemeyer, F. L. Carpenter, and J.
Kettler. 2001. Intercropping legume trees with native timber
trees rapidly restores cover to eroded tropical pasture without
fertilization. Forest Ecology and Management 152:195–209.
Orozoco Zamora, C., and F. Montagnini. 2007. Seed rain and
seed dispersal agents in pure and mixed plantations of native
trees and abandoned pastures at La Selva Biological Station,
Costa Rica. Restoration Ecology 15:453–461.
Otero-Arnaiz, A., S. Castillo, J. Meave, and G. Ibarra-
Manriquez. 1999. Isolated pasture trees and the vegetation
under their canopies in the Chiapas Coastal Plain, Mexico.
Biotropica 31:243–254.
Parrotta, J. A. 1992. The role of plantation forests in
rehabilitating degraded tropical ecosystems. Agriculture
Ecosystems and Environment 41:115–133.
Parrotta, J. A. 1995. Influence of overstory composition on
understory colonization by native species in plantations on a
degraded tropical site. Journal of Vegetation Science 6:627–
636.
Parrotta, J. A., O. H. Knowles, and J. M. Wunderle. 1997.
Development of floristic diversity in 10-year-old restoration
forests on a bauxite mined site in Amazonia. Forest Ecology
and Management 99:21–42.
Peh, K. S. H., N. S. Sodhi, J. de Jong, C. H. Sekercioglu,
C. A. M. Yap, and S. L. H. Lim. 2006. Conservation value of
degraded habitats for forest birds in southern peninsular
Malaysia. Diversity and Distributions 12:572–581.
Pejchar, L., R. M. Pringle, J. Ranganathan, J. R. Zook, G.
Duran, F. Oviedo, and G. C. Daily. 2008. Birds as agents of
seed dispersal in a human-dominated landscape in southern
Costa Rica. Biological Conservation 141:536–544.
Pennington, T. D., and E. C. M. Fernandes. 1998. The genus
Inga. Continental Printing, Brussels, Belgium.
Perfecto, I., R. A. Rice, R. Greenberg, and M. E. VanderVoort.
1996. Shade coffee: a disappearing refuge for biodiversity.
BioScience 46:598–608.
Peterson, C. J., and B. L. Haines. 2000. Early successional
patterns and potential facilitation of woody plant coloniza-
tion by rotting logs in premontane Costa Rican pastures.
Restoration Ecology 8:361–369.
Puyravaud, J. P., C. Dufour, and S. Aravajy. 2003. Rain forest
expansion mediated by successional processes in vegetation
thickets in the Western Ghats of India. Journal of
Biogeography 30:1067–1080.
Rickert, E. 2005. Environmental effects of the coffee crisis: a
case study of land use and avian communities in Agua Buena,
Costa Rica. Thesis. Evergreen State College, Olympia,
Washington, USA.
Robinson, G. R., and S. N. Handel. 2000. Directing spatial
patterns of recruitment during an experimental urban
woodland reclamation. Ecological Applications 10:174–188.
Rodrigues da Silva, U. D. S., and D. Matos. 2006. The invasion
of Pteridium aquilinum and the impoverishment of the seed
bank in fire prone areas of Brazilian Atlantic Forest.
Biodiversity and Conservation 15:3035–3043.
Russell-Smith, J., P. J. Stanton, P. J. Whitehead, and A.
Edwards. 2004. Rain forest invasion of eucalypt-dominated
woodland savanna, iron range, north-eastern Australia: I.
Successional processes. Journal of Biogeography 31:1293–
1303.
R. J. COLE ET AL.1268 Ecological Applications
Vol. 20, No. 5
Sanchez, P. A., C. A. Palm, C. B. Dabey, L. T. Szott, and C. E.
Russell. 1985. Tree crops as soil improvers in the humid
tropics? Pages 327–350 in M. G. R. Cannell and J. E.
Jackson, editors. Attributes of trees as crop plants. Institute
of Terrestrial Ecology, Natural Environmental Research
Council, Huntington, UK.
Schlawin, J., and R. A. Zahawi. 2008. ‘‘Nucleating’’ succession
in recovering neotropical wet forests: the legacy of remnant
trees. Journal of Vegetation Science 19:485–487.
Sekercioglu, C. H., S. R. Loarie, F. O. Brenes, P. R. Ehrlich,
and G. C. Daily. 2007. Persistence of forest birds in the Costa
Rican agricultural countryside. Conservation Biology 21:
482–494.
Slocum, M. G. 2000. Logs and fern patches as recruitment sites
in a tropical pasture. Restoration Ecology 8:408– 413.
Slocum, M. G. 2001. How tree species differ as recruitment foci
in a tropical pasture. Ecology 82:2547–2559.
Slocum, M. G., and C. C. Horvitz. 2000. Seed arrival under
different genera of trees in a neotropical pasture. Plant
Ecology 149:51–62.
Sodhi, N. S., L. P. Koh, D. M. Prawiradilaga, I. Tinulele, D. D.
Putra, and T. H. T. Tan. 2005. Land use and conservation
value for forest birds in Central Sulawesi (Indonesia).
Biological Conservation 122:547–558.
Systat Software. 2007. Systat 12.0. Systat Software, Inc.,
Chicago, Illinois, USA.
Thomlinson, J. R., M. I. Serrano, T. D. Lopez, T. M. Aide, and
J. K. Zimmerman. 1996. Land-use dynamics in a post-
agricultural Puerto Rican landscape (1936–1988). Biotropica
28:525–536.
Verdu´ , M., and P. Garcı
´a-Fayos. 1996. Nucleation processes in
a Mediterranean bird-dispersed plant. Functional Ecology
10:275–280.
Vieira, I. C. G., C. Uhl, and D. Nepstad. 1994. The role of the
shrub Cordia-multispicata Cham. as a succession facilitator in
an abandoned pasture, Paragominas, Amazonia. Vegetatio
115:91–99.
Wijdeven, S. M. J., and M. E. Kuzee. 2000. Seed availability as
a limiting factor in forest recovery processes in Costa Rica.
Restoration Ecology 8:414– 424.
Wright, S. J., and H. C. Muller-Landau. 2006. The future of
tropical forest species. Biotropica 38:287–301.
Wunderle, J. M. 1997. The role of animal seed dispersal in
accelerating native forest regeneration on degraded tropical
lands. Forest Ecology and Management 99:223–235.
Yarranton, G. A., and R. G. Morrison. 1974. Spatial dynamics
of a primary succession: nucleation. Journal of Ecology 62:
417–428.
Zahawi, R. A., and C. K. Augspurger. 2006. Tropical forest
restoration: tree islands as recruitment foci in degraded lands
of Honduras. Ecological Applications 16:464– 478.
Zanne, A. E., and C. A. Chapman. 2001. Expediting
reforestation in tropical grasslands: distance and isolation
from seed sources in plantations. Ecological Applications 11:
1610–1621.
Zar, J. H. 1996. Biostatistical analysis. Prentice Hall, Engle-
wood Cliffs, New Jersey, USA.
Zimmerman, J. K., J. B. Pascarella, and T. M. Aide. 2000.
Barriers to forest regeneration in an abandoned pasture in
Puerto Rico. Restoration Ecology 8:350–360.
APPENDIX A
Seed species collected in seed traps (Ecological Archives A020-044-A1).
APPENDIX B
Land use history and characteristics of research sites (Ecological Archives A020-044-A2).
APPENDIX C
A photo showing study sites in Coto Brus County in southern Costa Rica (Ecological Archives A020 -044-A3).
July 2010 1269SEED RAIN UNDER TREE ISLANDS
... These factors often resulted in recruitment limitation 8,87,[168][169][170] . Limitations due to seed dispersal-the most common form of seed dispersion in TMEs 159,171 -were exacerbated by habitat loss, fragmentation and degradation which disrupt seed dispersers' abundances and movement pathways 65,83,136,172,173 . Further, negative biotic interactions such as competition between grasses and ferns, pests and diseases, as well as herbivory and seed predation compromised restoration success. ...
... As part of this many studies specifically mentioned facilitation processes which ameliorate micro-environmental site conditions and often contributed to increased restoration success 54,104,174,175 . Facilitative interactions were deliberately employed in restoration studies, e.g. through applied nucleation tree island planting 11,172 , exotic plantations to recover native understories 129 , bracken ferns as facilitators for late succession tree seedlings 128 or planting to attract seed dispersers 120,140,141,172 . Moreover, site management variables related to removing disturbance, such as eradication of invasive species 84,176 , and protection of restoration sites 6,177 were mentioned as promoting success. ...
... As part of this many studies specifically mentioned facilitation processes which ameliorate micro-environmental site conditions and often contributed to increased restoration success 54,104,174,175 . Facilitative interactions were deliberately employed in restoration studies, e.g. through applied nucleation tree island planting 11,172 , exotic plantations to recover native understories 129 , bracken ferns as facilitators for late succession tree seedlings 128 or planting to attract seed dispersers 120,140,141,172 . Moreover, site management variables related to removing disturbance, such as eradication of invasive species 84,176 , and protection of restoration sites 6,177 were mentioned as promoting success. ...
Article
Full-text available
Many tropical mountain ecosystems (TME) are severely disturbed, requiring ecological restoration to recover biodiversity and ecosystem functions. However, the extent of restoration efforts across TMEs is not known due to the lack of syntheses on ecological restoration research. Here, based on a systematic review, we identify geographical and thematic research gaps, compare restoration interventions, and consolidate enabling factors and barriers of restoration success. We find that restoration research outside Latin-America, in non-forested ecosystems, and on socio-ecological questions is scarce. For most restoration interventions success is mixed and generally limited by dispersal and microhabitat conditions. Finally, we propose five directions for future research on tropical mountain restoration in the UN decade of restoration, ranging from scaling up restoration across mountain ranges, investigating restoration in mountain grasslands, to incorporating socio-economic and technological dimensions.
... These factors often resulted in recruitment limitation 8,87,[168][169][170] . Limitations due to seed dispersal-the most common form of seed dispersion in TMEs 159,171 -were exacerbated by habitat loss, fragmentation and degradation which disrupt seed dispersers' abundances and movement pathways 65,83,136,172,173 . Further, negative biotic interactions such as competition between grasses and ferns, pests and diseases, as well as herbivory and seed predation compromised restoration success. ...
... As part of this many studies specifically mentioned facilitation processes which ameliorate micro-environmental site conditions and often contributed to increased restoration success 54,104,174,175 . Facilitative interactions were deliberately employed in restoration studies, e.g. through applied nucleation tree island planting 11,172 , exotic plantations to recover native understories 129 , bracken ferns as facilitators for late succession tree seedlings 128 or planting to attract seed dispersers 120,140,141,172 . Moreover, site management variables related to removing disturbance, such as eradication of invasive species 84,176 , and protection of restoration sites 6,177 were mentioned as promoting success. ...
... As part of this many studies specifically mentioned facilitation processes which ameliorate micro-environmental site conditions and often contributed to increased restoration success 54,104,174,175 . Facilitative interactions were deliberately employed in restoration studies, e.g. through applied nucleation tree island planting 11,172 , exotic plantations to recover native understories 129 , bracken ferns as facilitators for late succession tree seedlings 128 or planting to attract seed dispersers 120,140,141,172 . Moreover, site management variables related to removing disturbance, such as eradication of invasive species 84,176 , and protection of restoration sites 6,177 were mentioned as promoting success. ...
Article
Full-text available
Many tropical mountain ecosystems (TME) are severely disturbed, requiring ecological restoration to recover biodiversity and ecosystem functions. However, the extent of restoration efforts across TMEs is not known due to the lack of syntheses on ecological restoration research. Here, based on a systematic review, we identify geographical and thematic research gaps, compare restoration interventions, and consolidate enabling factors and barriers of restoration success. We find that restoration research outside Latin-America, in non-forested ecosystems, and on socio-ecological questions is scarce. For most restoration interventions success is mixed and generally limited by dispersal and microhabitat conditions. Finally, we propose five directions for future research on tropical mountain restoration in the UN decade of restoration, ranging from scaling up restoration across mountain ranges, investigating restoration in mountain grasslands, to incorporating socioeconomic and technological dimensions. Tropical mountain ecosystems (TME) are hotspots of biodiversity 1,2 and endemism 3 and are located in tropical latitudes between 1000 and 4000 m asl, and the elevation gradients give rise to a variety of ecosystems including montane forests, montane cloud forests, forest-grassland treelines, mountain grasslands and azonal formations (Table 1). TME span across all continents in the tropical belt, and despite their small spatial extent of just over 4 million km 2 (Table 1) they provide numerous ecosystem services to people and society, including carbon sequestration, water regulation and supply, timber and food provision, erosion control, and cultural services 4. Notwithstanding their tremendous biological importance and complexity, TME are still relatively under-studied compared to temperate mountain systems 5. In recent decades, TME have been experiencing increasing pressure from multiple external drivers and stressors, such as anthropogenic pressures due to agricultural encroachment, pasture conversion and population growth 6 , exotic plantations 7,8 , invasion by exotic animals 9 and exotic plants 10-12 , as well as accelerating climate change impacts 13. These drivers lead to severe degradation in TME, impacting all levels of ecological organization, such as disruption of ecosystem services, losses in community diversity, changes in species interactions, reductions of population sizes and lowered genetic diversity 14. Degradation in TME is far-reaching and ubiquitous: Tovar et al. 15 projected that climate change will alter 3-7% of tropical Andean biomes, resulting in a 31.4% loss in extent of high-altitudinal Páramo grasslands due to replacement by montane forests by 2039. Further, Helmer et al. 16 indicate that in the next 25-45 years, reductions in cloud immersion are estimated to diminish 57-80% of Neotropical montane cloud forests. Hall et al. 17 estimate that the Tanzanian Eastern Arc mountains have lost 25% of forested areas since 1955, with deforestation rates of 57% in sub-montane forests (800-1200 m). At the same time, socioeconomic drivers have led to migrations of people from tropical mountains to urban areas, abandoning many previously cultivated and inhabited areas 24-26 and creating a large opportunity for ecosystem recovery and restoration across many TME. Restoration of biodiverse ecosystems, such as TME, has the potential to simultaneously recover lost biodiversity and ecosystem functioning and improve local livelihoods 27 , and has recently come to the fore of global conservation efforts 28. Restoration is defined as "the process of assisting the recovery of an ecosystem that has been degraded, damaged or destroyed" 29 and, as such, encompasses a broad suite of approaches ranging from passive restoration, to assisted recovery and active restoration. The urgency for global restorative actions culminated in global restoration pledges like the 2011 Bonn Challenge and the proclamation of the UN Decade of Ecosystem Restoration. Motivations to restore damaged ecosystems include conserving biodiversity (specific habitats or species), enhancing ecosystem processes (such as nutrient cycling), combatting climate change (through carbon OPEN
... The normal recovery of mining-degraded sites takes time for plant and animal species to colonize [5][6][7]. However, in many degraded areas, conventional recovery practices combined with biological restoration through human involvement can speed up the restoration process [8,9]. ...
... Total number of quadrats in which the species occurred (6) In both FA and DA abundance of herbaceous species was estimated at regular intervals and categorized as rare, common and abundant. Further abundance rating (AR) was given to each herbaceous species. ...
... Furthermore, the study suggests that growing leguminous plants, which give nitrogen to the root zone in mine spoil, would be more beneficial in increasing mining wasteland fertility [19]; [33]. Albizia lebbeck, Acacia auriculiformis and Acacia nilotica are the species with strong nitrogenous activity in root nodules [6]; [34]. As a result, these tree species are recommended for planting on dumps and in degraded environments. ...
Chapter
Full-text available
Mining activities in Jajang iron and manganese ore mines located in Keonjhar district of Odisha, India starting from mineral explorations to production and transport are causing environmental damage in many ways, which includes deforestation , loss of topsoil, accelerated soil erosion, migration of wildlife and avifauna, and addition of air pollutants and dust to the atmosphere. In connection to this, the current study was an attempt to regain the original ecological status of the degraded areas of Jajang iron and manganese ore mines caused due to mining by Rungta Mines Limited. To achieve this indigenous plant species for restoration were selected from mining forests and plantations. Species selection from mining forests was made through systematic phytosociological analysis that involved measurement of Importance Value Index (IVI), regeneration values of tree species and their economic uses. On the other hand, species selection from plantations was made based on their growth, productivity, economic uses and adaptation to terrain and soil types. Shrubs and grasses were selected based on their relative index and abundance, respectively. The top 15 tree and 16 grass species as well as all six shrub species were selected from mining forests and plantations were considered for restoration. The findings of the study may also aids in the faster restoration of degraded habitats with initial human facilitation as the soils of degraded areas were similar to that of the mining forest. To speed up the recovery process after-care and monitoring have also been suggested or advised.
... Posteriormente, este modelo ha recibido diferentes etiquetas en la literatura científica, tales como islas de árboles (Cole et al. 2010), nucleación aplicada (Corbin y Holl 2012) y núcleos de dispersión y reclamo (García-Martí y Ferrer 2013). Otro esquema relacionado es el de las plantaciones forestales ubicadas en las esquinas de campos agrícolas (Hughes-Clarke y Mason 1992). ...
... Posteriormente, este modelo ha recibido diferentes etiquetas en la literatura científica, tales como islas de árboles (Cole et al. 2010), nucleación aplicada (Corbin y Holl 2012) y núcleos de dispersión y reclamo (García-Martí y Ferrer 2013). Otro esquema relacionado es el de las plantaciones forestales ubicadas en las esquinas de campos agrícolas (Hughes-Clarke y Mason 1992). ...
... Posteriormente, este modelo ha recibido diferentes etiquetas en la literatura científica, tales como islas de árboles (Cole et al. 2010), nucleación aplicada (Corbin y Holl 2012) y núcleos de dispersión y reclamo (García-Martí y Ferrer 2013). Otro esquema relacionado es el de las plantaciones forestales ubicadas en las esquinas de campos agrícolas (Hughes-Clarke y Mason 1992). ...
... The presentation focused specifically on three of these mechanisms: autocatalytic nucleation (defined above, Michaels et al. 2020), directed dispersal, and resource concentration. Here, directed dispersal refers to patches of trees and shrubs increasing the local seed rain, by attracting dispersers in the surrounding landscape toward these patches (Cole et al. 2010, Caughlin et al. 2016. Resource concentration refers to organisms within a patch harvesting resources from their surroundings, improving growth conditions within the patch at the expense of resource availability at larger spatial scales (Rietkerk et al. 2004, Rietkerk andvan de Koppel 2008). ...
... The central argument that remnant forest habitat is important to the natural recovery of adjacent degraded areas (Crouzeilles et al. 2020) is not disputed in the literature. However, a number of studies have found a minimal effect of the amount of remaining forest in a given area on recovery patterns, including in this experimental study (Cole et al. 2010, Holl et al. 2017, creating a seeming paradox with major potential implications for conservation and restoration efforts. We hypothesized that such results may be an artifact of the coarse resolution at which remnant forest cover is assessed, and predicted that distinct patterns would emerge if recruit dynamics were examined at the species level by mapping the presence of adult trees in remnant forest habitat surrounding targeted restoration plots. ...
Article
Full-text available
Remnant trees and forest fragments in agricultural landscapes can be important sources of propagules to facilitate forest recovery. However, many studies simply quantify forest cover in the surrounding landscape as a percentage, with little attention given to species composition, and subsequently fail to detect an effect on recruitment patterns. We assessed the relative importance of the spatial distribution and life-history traits of 77 tree species on recruitment patterns at a landscape scale in a well-replicated long-term restoration study in southern Costa Rica. We censused and mapped potential mother trees in a 100-m buffer surrounding eight replicate restoration plots and quantified respective tree recruits within each plot. We assessed how mother tree abundance, species life-history characteristics (seed size, dispersal mode), tree size (DBH, height) and distance to restoration plot affected recruitment at coarse (plot: 50 × 50 m) and fine (quadrat: 3 × 3 m) spatial scales. The presence of a mother tree within 100 m of a restoration plot resulted in a 10-fold increase in potential mean recruitment. Mother tree abundance was also an important driver of recruit density, and particularly so for large-seeded (≥ 5 mm) zoochorous species with a fivefold increase in recruit density across the observed mother tree abundance range. An interaction between mother tree abundance and proximity demonstrated that the effect of mother tree abundance on recruit density was important but waned with increasing distance from restoration plots. At the fine spatial scale, proximity was uniformly important; height and DBH of the closest potential mother tree also affected recruit abundance but responses differed by seed size. Results highlight the importance of remnant vegetation composition to the recovery of adjacent degraded habitats, underscoring the outsized role nearby remnant forest and isolated trees can play for the persistence of localized biodiversity.
... Studies of this phenomenon are common in forest ecosystems [6], especially in tropical rainforests [7]. Studies of anthropically impacted landscapes are also numerous [3,8,9,10], but data on subsidence basins, a unique habitat, are missing. The aim of this work is therefore to provide basic information, such as composition, density, and similarity to above-ground vegetation, about the seed rain of these habitats. ...
Article
Full-text available
Overall amount and quality of seeds reaching the area is an important factor in the development of plant communities in each habitat. It can be a significant source of diversity in anthropogenically disturbed areas. This study is focused on the survey of seed rain in habitats of waterlogged subsidence basins of the Karviná region, induced by deformations of overlying rocks in the area of deep coal mining. The density, species composition and dispersal strategies of captured seeds were monitored using funnel traps with a 35 cm diameter, fixed at a height of 0.5 m. A vegetation study was conducted to compare seed rain with above-ground vegetation using the Jaccard similarity index. Over 11 000 seeds of 27 plant species were captured in total. Seed rain was dominated by Salix alba species. Seed density varied significantly within individual sampling plots (7–12380 seeds/m2) with the average richness of four species. The diversity of above-ground vegetation was also low, with an average of 15 species per plot (100 m2). Similarity between seed rain and above-ground vegetation of subsidence basins is 27.6 % and 52 % within the woody species community. It is evident that the seed rain of waterlogged subsidence basins contains species of floodplain forests and wetland vegetation. On the other hand, invasive and expansive species identified both in seed rain and above-ground vegetation can potentially endanger the development of the succession of these habitats. The analysis of seed dispersal also shows a significant proportion of species of arid and mesic habitats.
... Enrichment planting makes the oil palm landscape more accessible to native species to form clusters of tree islands within the plantation Phalan et al., 2009). These tree islands in agricultural landscapes are a cost-effective measure to enhance biodiversity and ecosystem functions (Cole et al., 2010). ...
Article
About 2.5 million hectares of a total of 15 million hectares of oil palm plantation in Indonesia are planted in, or conflict with, the forest zone. Oil palm plantations face a conflict between socio-economic and ecological issues. This study was conducted in the Harapan Rainforest, Jambi to evaluate the potential of oil palm-based agroforestry to reconcile economic and ecological interests, by considering socio-economic and financial feasibility as well as biodiversity and land cover. The financial feasibility of oil palm agroforestry is compared to oil palm monoculture, employing a discounted cash flow approach using three indicators: net present value (NPV), benefit-cost ratio (BCR), and internal rate of return (IRR). Two ecological indicators—biodiversity and land cover—are evaluated in an experimental plot of oil palm agroforestry in Jambi. This study indicates that the NPV, BCR, and IRR of oil palm monoculture are IDR 62,644,836 (US$ 4,476.84), 1.39, and 20.77%, respectively, while the oil palm agroforestry planted in the experimental plot potentially generates much better values of financial indicators with NPV, BCR, and IRR being IDR 209,221,212 (US$ 14,951.76), 1.79, and 24.42%, respectively. Besides evaluating financial feasibility, we also found that the reviewed current studies indicate that the oil palm agroforestry provides positive ecological impacts, such as increased forest land cover, invertebrate fauna, and bird diversity.
Article
Full-text available
Planted agricultural windbreaks may provide habitat for forest tree species and facilitate forest regeneration within the agricultural landscape, if the windbreaks are colonized by forest trees. In order to examine the potential for windbreaks to foster forest regeneration, I surveyed the understories of 51 windbreaks in Monteverde, Costa Rica, for forest tree seedlings. The windbreaks had been planted 5-6 yr earlier. Ninety-one forest tree species (including both primary and secondary forest species) were found in the wind-breaks; however, many of the species occurred in low densities. The mean density of forest tree seedlings in the windbreaks was 2.14 seedlings/m2. Landscape factors appeared to have important effects on tree recruitment patterns. Wind-breaks that were connected to forest had significantly higher densities of forest tree seedlings and greater numbers of forest tree species in their understories than windbreaks that were not connected to forest. In particular, connected windbreaks had higher densities and more species of bird-dispersed trees than nonconnected windbreaks. These differences probably reflect the greater activity of frugivorous birds within connected windbreaks and the resulting higher input of forest tree seeds. Windbreaks with remnant forest trees had significantly higher densities of forest tree seedlings of other species and more forest tree species than windbreaks lacking remnant trees. By serving as habitats for forest tree species, windbreaks may contribute to the maintenance of forest trees within agricultural landscapes and serve as tools for forest restoration in the tropics.
Article
Full-text available
Patterns of plant and animal diversity were studied in a 10-year-old native species reforestation area at a bauxite-mined site at Porto Trombetas in western Para State, Brazil. Understorey and overstorey floristic composition and structure, understorey light conditions, forest floor development and soil properties were evaluated in a total of 38 78.5-m 2 plots located in the reforestation area at varying distances up to 640 m from the boundary with the surrounding primary forest. Wildlife surveys focusing primarily on birds and bats were also conducted to assess the role of seed-dispersing animals in regeneration of woody forest species within the plantations and colonization by primary forest species not included in the original reforesta­ tion. Regeneration density, species richness and species diversity (Shannon-Wiener index) for woody perennial species, vines, herbs and grasses were strongly correlated with the diversity of planted tree species and structural development parameters, degree of forest floor development and soil pH. The better developed closed-canopy plots (>40% crown closure) were characterized by relatively well-developed litter (01) and humus (02) layers, more acidic soils being typical of the surrounding primary forests, and a more diverse herb, vine and woody perennial flora with a greater representation of primary forest species being characteristic of late secondary forests. In closed-canopy plots a total of 125 tree, palm and shrub species were censused (versus 34 in the more open-canopy plots), of which 75 species are known to have been introduced by natural means from the surrounding primary forest (versus 1 1 species in the open-canopy plots). Among the plantation plots, there was significant colonization by primary forest woody species up to 640 m away from the primary forest edge, although both the abundance and the diversity of colonizing species declined with increasing distance into the plantations. Smaller-seeded primary forest woody species dispersed by mammals and birds represented a higher proportion of the colonizing species compared with the larger-seeded species. These data are consistent with the results of the wildlife surveys, which indicated that most animal seed dispersal is provided by bats, that the most common frugivorous bat and bird species in the plantations feed on small-seeded plant species and that birds and mammals that typically disperse larger-seeded tree species (such as toucans, trogons, tapirs, deer and primates) are still rare in the reforestation area. These results suggest that while the reforestation program has been successful in creating a favorable environment for regeneration of a native primary forest species, further management interventions, such as enrichment plantings, may be required to accelerate regeneration of large­ seeded primary forest species.
Article
Full-text available
The Coastal Plain of Chiapas (southern Mexico) was formerly covered by large tracts of subhumid tropical forests but is heavily deforested at present. In this region, 15 pastures were selected to characterize species composition of isolated trees, as well as to describe species composition, growth form patterns, and patterns of dispersal units in the vegetation growing under their canopies. The 65 recorded pasture trees belonged to 20 species and 11 families, of which Fabaceae and Moraceae were the most species-rich. Coccoloba barbadensis (Polygonaceae) and Enterolobium cyclocarpum (Fabaceae) were the most abundant and frequent trees in the studied pastures. More than half (55%) of isolated tree species were fleshy-fruited. In the vegetation sampled under pasture trees, 134 species and 45 families were found. Fabaceae and Poaceae had the largest numbers of species. Herbs were the predominant growth form (46.3%), followed by shrubs (23.9%), trees (23.1%), and lianas (2.2%); 6 species could not be placed in any growth form category. Most species of this flora were fleshy-fruited (43.3%), followed by heavy, gravity-dispersed fruits (17.9%). The analysis of dispersal units by growth form category confirmed the prevalence of fleshy fruits, although their predominance was not so obvious among herb species. Almost half (49.2%) of the flora under pasture trees was typical of secondary vegetation; this pattern was true for herbs but not for most woody species, which were typical of primary vegetation. A numeric classification of the vegetation samples taken under pasture trees produced eight floristic groups, all of which were independent of the specific identity of pasture trees. No significant effect of dispersal unit type of pasture tree on the characteristics of the vegetation growing under them was found. Future attempts to re-create the original forest cover using isolated trees in pastures as regeneration foci should pay more attention to the maintenance of a large specific diversity independently of the dispersal types among these components of tropical landscapes.
Article
Full-text available
Studies of biological invasions indicate that natural recruitment of new species can occur as a 'nucleation' phenomenon, in which scattered colonization foci spread and coalesce. Ecological reclamation of damaged lands might make use of this potential for enhanced natural dispersal, by inoculating sites with multiple small plantings to attract animal dispersers and other mutualists from nearby remnants of natural habitat. We conducted an experimental test of this proposition. On a 6-ha section of an abandoned municipal landfill in the New Jersey Meadowlands, we installed 16 clusters of 21 trees and shrubs in an array of fenced plots. Clusters contained seven native species known to: (1) attract bird dispersers to introduce propagules from remnants of off-site habitat; (2) contribute propagules by virtue of high reproductive output and clonal growth; and (3) accelerate woodland succession on open, degraded habitats. Average plant size was varied, with half the plots receiving larger trees and shrubs, to test whether woody plant size would enhance any attractive function. An additional eight empty plots were studied to estimate background rates of recruitment and to test for a fencing effect. Site preparation included the addition of 90 cm of fresh substrate, including organic matter, and a cover crop of annual grasses. Recruitment of woody plants inside and surrounding the experimental plots was examined for five years, and results were compared on the basis of treatment and recruitment mode (avian, wind, or clonal dispersal). Woody plant recruitment into experimental plots was rapid and substantial, primarily via dispersal from natural sources. Plots with larger plants attracted significantly more recruits at the outset, but this difference diminished over time. Fall seed rain samples yielded a mean estimate of 426 seeds/m2 within plots. However, size distributions of recruiting woody species increasingly shifted toward larger individuals each year. Experimental manipulations that opened seed beds for woody plant recruitment had short-lived effects, indicating a narrow window of opportunity for establishment. Spread of the planted species themselves was generally weak, although clonal growth contributed substantially to spread on the margins of plots. Most recruitment outside experimental plots was from external sources. A strong proximity component was found for bird-dispersed recruits, which were highly clustered near planted plots, with the highest densities near source populations on the site margin. Wind-dispersed trees and shrubs, by contrast, were not associated with planted plots and were concentrated near one corner of the site. Discounting plot interiors, total recruitment density for the site after 5 yr was ~800 woody stems/ha, 36% via avian dispersal, 10% via clonal spread, and the remainder via wind-borne propagules. New recruits represented 26 woody plant species, all but four from external sources, and only five common species contributed more than a few recruits. We conclude that techniques for manipulating natural seed dispersal hold promise for ecological restoration, provided that background populations are available to supply colonists.
Article
Anthropogenic alteration of the landscape is a long-term disturbance both in duration and consequences. This study addresses land-cover responses to a history of human land use in northeastern Puerto Rico. Analysis of aerial photographs indicated that the pattern of land use in the municipality of Luquillo changed dramatically between 1936 and 1988. In 1936, sugar cane and pasture were the dominant land uses, occupying about one third of the study area each, while dense forest was rare. Pasture still occupied about a quarter of the area by 1988, but the area of sugar cane had declined to zero. Most sugar cane land was transformed to pasture after abandonment, while much of the pasture at higher elevations reverted to forest. More than half the study area in 1988 was occupied by dense forest, and the degree of forest regeneration was greatest adjacent to the Luquillo Experimental Forest and around parches that were dense forest remnants in 1936. The overall trend was from high-intensity agriculture to dense forest, bur urban areas increased more than 2000 percent between 1936 and 1988 and are presently encroaching on forested areas. It is unclear from one study whether the same pattern would hold true at other sites in the tropics, but our study indicates the importance of preserving remnants of mature forest as sources of forest regeneration. In addition, the nature of the patches in the Luquillo landscape has changed as the land use has changed. In 1936, dense forest was highly fragmented, the patches were small and many of them had linear configurations (riparian corridors and hedgerows). By 1988, the average parch size of dense forest had increased greatly, although with one exception the parches were still small relative to other land-cover types.
Article
The 1st woody species to appear in plots after felling are light-demanding species growing from seeds remaining dormant in the rain forest soil. Other seeds are disseminated by frugivorous bats and birds, yet others by wind. A few other regenerating plants are sucker shoots from cut stumps of shade tolerant species of the undergrowth. The pioneer vegetation is characterized by a low species richness: no more than 15 species (out of a total of 113 found 3.5 yr after clear cutting), contribute to 85% of the woody community. The seeds of these pioneer species are dispersed by bats and birds. Twenty species of frugivorous bats were noted in the area, of which only 4 were abundant in the regenerating forest. Forty species of birds also contribute to seed dispersal, of which 16 species are restricted to second-growth. Some of these birds have a limited flight range and tend to disperse seeds more or less evenly around feeding trees. Other birds, particularly lek forming species (Pipridae), tend to defecate mainly around their traditional display grounds and generate a more aggregative dispersal of seeds. Some generalist species with a mixed insectivorous and frugivorous diet (Turdus albicollis and Ramphocelus carbo) are the first birds to colonize newly formed gaps in the forest. -from Authors
Article
Ranching and logging operations are transforming the moist tropical forests of an eastern Amazonian landscape into a mosaic of pastures and regrowth forests. The new ecosystems of this region are agriculturally unproductive, biologically impoverished, and far more flammable than the mature forests they replace. In the absence of fire, the forest regrows on abandoned sites, accumulating biomass and species at a rate that is inversely related to the intensity of use prior to abandonment. Forest regrows slowest on those rare abandoned pastures that were once scraped with bulldozers. The grass- and shrub-dominated old fields that form on some of these sites resist forest regrowth because of numerous barriers to tree establishment and growth. Knowledge of these barriers provides a basis for developing inexpensive techniques to restore agricultural productivity in old fields by implanting tree-based agricultural systems or to restore forest regenerative capacity in old fields by establishing trees that attract seed-carrying animals and ameliorate harsh environmental conditions. These restoration techniques will be needed over large areas of Amazonia if current attempts to reform degraded pastures fail. -from Authors