Ecological Applications, 26(1), 2016, pp.

203218
2016 by the Ecological Society of America

Acceleration and novelty: community restoration speeds recovery

and transforms species composition in Andean cloud forest
SARAH JANE WILSON1,3 AND JEANINE M. RHEMTULLA2
1Department of Geography, McGill University, 805 Sherbrooke Street West, Montreal, Quebec H3A 0B9 Canada
2Department of Forest and Conservation Sciences, University of British Columbia, 2424 Main Mall, Vancouver, British
Columbia V6T 1Z4 Canada

Abstract. Community-based tropical forest restoration projects, often promoted as a

win-win solution for local communities and the environment, have increased dramatically
in number in the past decade. Many such projects are underway in Andean cloud forests,
which, given their extremely high biodiversity and history of extensive clearing, are
understudied. This study investigates the efficacy of community-based tree-planting projects
to accelerate cloud forest recovery, as compared to unassisted natural regeneration. This
study takes place in northwest Andean Ecuador, where the majority of the original, highly
diverse cloud forests have been cleared, in five communities that initiated tree-planting
projects to restore forests in 2003. In 2011, we identified tree species along transects in
planted forests (n = 5), naturally regenerating forests (n = 5), and primary forests (n =
5). We also surveyed 120 households about their restoration methods, tree preferences,
and forest uses. We found that tree diversity was higher in planted than in unplanted
secondary forest, but both were less diverse than primary forests. Ordination analysis
showed that all three forests had distinct species compositions, although planted forests
shared more species with primary forests than did unplanted forests. Planted forests also
contained more animal-dispersed species in both the planted canopy and in the unplanted,
regenerating understory than unplanted forests, and contained the highest proportion of
species with use value for local people. While restoring forest increased biodiversity and
accelerated forest recovery, restored forests may also represent novel ecosystems that are
distinct from the regions previous ecosystems and, given their usefulness to people, are
likely to be more common in the future.
Key words: Andes; biodiversity; cloud forest; community forestry; forest use; locally useful species; novel
forests; reforestation; restoration; succession; tropical forest; watershed management

resources while improving local environmental condi-

INTRODUCTION
In response to rising concerns about biodiversity
and species loss in tropical forests, NGOs, govern-
ments, and other organizations increasingly encourage
local communities to restore forest in degraded tropical
landscapes. From an ecological perspective, the aim
of tropical forest restoration is to accelerate or jump-
start succession, helping forests recover lost ecosystem
functions and species diversity more quickly than if
left to regenerate unassisted (Lamb et al. 2005, Harris
and van Diggelen 2006). At the same time, develop-
ment agencies promote community-based restoration
to provide local people with employment and forest
Manuscript received 10 November 2014; revised 19 March
2015; accepted 25 March 2015. Corresponding Editor:
E. Cienciala.
3 E-mail: sarah.wilson@mail.mcgill.ca
tions (i.e., the United Nations Billion Tree Campaign).
The win-win potential of these projects has made
community-based restoration, which communities
actively plan and manage (Charnley and Poe 2007,
Agrawal and Angelsen 2009), common in many tropi-
cal regions (e.g., Kumar et al. 2005). But how well
these projects work to achieve ecological goals, includ-
ing biodiversity conservation, is still poorly understood.
This study asks: can community-based restoration
projects work to restore biodiverse cloud forest, or
do the trade-offs between social and ecological goals
outweigh the synergies? (Wunder 2007:49, Adams
et al. 2004, Chhatre and Agrawal 2008).
Ecological forest restoration aims to overcome bar-
riers to succession in degraded landscapes, ultimately
creating forests that are similar in function, structure,
or both, to the primary forests that once occupied

203
204 S. J. WILSON AND J. M. RHEMTULLA
the site (Lamb et al. 2005, Chazdon 2008). Restoration
is especially important where natural regeneration is
slow or impeded (Chazdon 2003, 2008). The likelihood
and speed of natural tropical forest regeneration are
affected by: the intensity and duration of previous
land use; proximity to mature forests; and the pres-
ence in the soil of propagules (e.g., seed banks and
sprouting stumps and roots), seed dispersers, and rem-
nant vegetation (Uhl et al. 1988, Parrotta et al. 1997,
Harvey 2000, Holl et al. 2000, Guariguata and Ostertag
2001, Chazdon 2003, Chazdon et al. 2003, Florentine
and Westbrooke 2004). Although biomass can recover
quickly in secondary forests, tree composition may
remain distinct for decades (Letcher and Chazdon 2009,
Bonner et al. 2013); in the southern Atlantic rain
forests in Brazil, for example, after 50 years, second-
ary forests shared only 19% of the same species with
nearby primary forest (Liebsch et al. 2008). Forest
recovery is often particularly slow on pastures that
have been heavily used for long periods, which now
occupy large areas of Latin America (Aide and Cavelier
1994, Ellis and Ramankutty 2008). Those forests that
do regenerate in these conditions may support different
plant species than primary forest for decades to cen-
turies (Uhl et al. 1988, Lamb et al. 2005, Sampaio
et al. 2007, Dent and Wright 2009). Planting trees on
such degraded lands can alter conditions so as to allow
other species to arrive and establish, creating forests
that, at least in the short term, are more species rich
than if they were allowed to regenerate naturally
(Parrotta 1993, 1995, Brockerhoff et al. 2008, Holl
and Aide 2011, de la Pea-Domene et al. 2013).
Putting local people in charge of managing local
resources, including restoring forests, is increasingly
common (Lamb et al. 2005, Agrawal et al. 2008).
Community-based forest management and restoration
are often promoted over top-down conservation
approaches because local people are thought to be
better positioned to monitor and manage forests than
outsiders (Agrawal and Angelsen 2009) and stand to
benefit from increased local forest cover and ecosystem
services. But, because social and ecological goals may
differ between ecologists and locals, and even among
local people within the same community (Agrawal and
Gibson 1999, Manzi and Coomes 2009), the species
that people prefer to plant might also differ from
those chosen for ecologically oriented restoration pro-
jects. For example, local people may choose to plant
exotic species common in cultural landscapes, or select
species based on their use value or familiarity (Michon
et al. 2007, Garen et al. 2009), thus potentially under-
mining the historical fidelity and/or ecological integrity
of the restoration by introducing nonnative species
(Higgs 2003, Hobbs et al. 2009).
How effectively community-based tree planting helps
to restore ecological characteristics in tropical forests is
not well studied, especially for montane cloud forests
(Aide et al. 2010). Cloud forests are rare, highly
Ecological Applications
Vol. 26, No. 1
biodiverse, and undergoing rapid conversion (Myers et al.
2000, Scatena et al. 2010). Comprising only 1.1% of
tropical forests in Latin America, over 50% of montane
cloud forest has been cleared, largely for pasture (Gentry
1989, Henderson et al. 1991, Jokisch and Lair 2002,
Wassenaar et al. 2007, Mulligan 2010). These forests
contain high numbers of endemic species, account for
up to 10% of bird and amphibian species globally, and
play a unique role in the hydrological cycle by capturing
mist from passing clouds (Stattersfield et al. 1998). But
because of their relatively thin soils, steep slopes, cool
temperatures and, in many regions, long history of land
use, high degree of fragmentation, and conversion to
pasture planted with aggressive pasture grasses, cloud
forests are particularly slow to recover after clearing
(Aide and Cavelier 1994, Churchill et al. 1995, Baillie
1996, Rhoades et al. 1998, Aide et al. 2010, Ortega-Pieck
et al. 2011). For example, even 15 years after abandon-
ment, Aide and Cavelier (1994) found little natural
regeneration on pastures planted with exotic grasses. Land
abandonment rates are currently high in some Andean
regions, providing an opportunity for reforestation on
former agricultural lands (Portillo-Quintero et al. 2012).
Effective restoration is thus a vital tool to increase forest
cover (Sarmiento 1995a, b, Young and Leon 1995), and
many community-based restoration projects are underway
in the northern Andes (Berry and Clunas 2009; informa-
tion on Maquipucuna Foundation available online).4
Despite this, we are not aware of other multi-site studies
that focus explicitly on biodiversity recovery in community-
restored forests, nor that compare the diversity of restored
forests in Andean montane cloud forest to spontaneous
forest regeneration.
In this study, we ask: does community-based tree
planting accelerate cloud forest recovery? Specifically,
compared to naturally regenerating secondary forests:
(1) Does planting trees increase tree species richness in
secondary forest? (2) Is the species composition of planted
forests overall more similar to that of primary forests?
(3) Are naturally regenerating (unplanted) understory
seedlings and saplings more similar in composition to
primary forests (i.e., is there evidence that planting trees
is accelerating succession)? (4) What species do people
choose to plant, and are these choices compatible with
the goal of restoring biodiverse cloud forests? We herein
address these questions by comparing primary, planted
secondary, and unplanted secondary forests in a multi-
site study in the cloud forests of Andean Ecuador.
METHODS
Study site
Our study was conducted in northwest Andean
Ecuador, part of the Tropical Andes biodiversity hot-
spot, in five communally reforested reserves all located
4 http://www.maqui.org/
January 2016 COMMUNITY-BASED CLOUD FOREST RESTORATION 205

TABLE 1. Description of planted, unplanted, and primary forest sites studied.

Site ID Elevation Land use history Reserve Surrounding land Year No. spp. No. spp.
(masl) size (ha) use abandoned planted IDed
Apuela
1 UP 20002050 cleared 1970s, crops (4 19 pasture, 2004 0 10
yr), pasture (30 yr) secondary forest
1 PL 18801920 cleared 1970s, crops (4 19 pasture, 2004 12 34
yr), pasture (30 yr) secondary forest
El Crystal
2 UP 21402200 cleared 1970s, crops (4 25.5 pasture, crops, 2004 0 13
yr), pasture (30 yr) secondary forest
2 PL 21002150 cleared 1970s, crops (4 25.5 pasture, crops, 2004 27 35
yr), pasture (30 yr) secondary forest
El Paraiso
3 UP 20002100 cleared 1970s, crops (4 40 pasture, crops 2006 0 17
yr), pasture (25 yr)
3 PL 20002100 cleared 1970s, crops (4 40 pasture, crops 2006 33 37
yr), pasture (28 yr)
La Esperanza
4 UP 21702190 cleared 1970s, crops (4 8.5 secondary forest, 2003 0 19
yr), pasture rotated pasture, crops
with crops (30 yr)
4 PL 21402165 cleared 1970s, crops 8.5 secondary forest, 2003 19 40
(4 yr), pasture pasture, crops
rotated with crops
(30 yr)
Pueblo Viejo
5 UP 20502080 cleared 1970s, crops 13 secondary forest, 2003 0 29
(4 yr), pasture pasture, crops
rotated with crops
(30 yr)
5 PL 2040 cleared 1970s, crops (4 13 secondary forest, 2002 18 75
yr), pasture rotated pasture, crops
with crops (30 yr)
Bosque Intag
6 19602060 primary forest 730 pasture, na 0 81
secondary and
primary forest
Junin
7 20402150 primary forest 5709 secondary forest, na 0 120
primary forest,
pasture
Los Cedros
8 19502100 primary forest 6880 secondary forest, na 0 104
primary forest,
pasture
Nangulbi Alto
9 19802000 primary forest 30 secondary forest, na 0 78
crops, pasture
Santa Lucia
10 20002150 primary forest 730 primary forest, na 0 58
secondary
forest, pasture

Notes: ID is the number of each site used in Figs. 24. Sites marked UP are unplanted secondary forest, PL are planted second-
ary forest, and not marked are primary. Primary forest refers to forest that has not been cleared in the past 80 years or more.
Elevation is measured in meters above sea level (masl). Surrounding land use refers to areas bordering the reserve (based on
S. Wilson, personal observation). The order in which they are listed corresponds to the relative proportion of the reserve sur-
rounded by each. Abandoned refers to the year that land in watershed reserves was planted with trees (planted sites) or left to
regenerate naturally (unplanted sites). Number of species IDed refers to the number of species found at each site.
206 S. J. WILSON AND J. M. RHEMTULLA
between 1800 and 2250 m above sea level (masl) in
the Intag Valley (Imbabura province, Ecuador; 0.35
N, 78.5 W; Fig. 1). The region is mountainous and
steep, with a mean annual temperature of 1720C,
annual rainfall of 15003300 mm, and a pronounced
dry season from May/June to October (Freiberg and
Freiberg 2000, Berry and Clunas 2009). The area is
characterized by exceptionally high levels of plant spe-
cies richness and endemism (Myers et al. 2000,
Sarmiento 2002); primary cloud forests here house over
300 species of trees per hectare, or up to 120 species
in a 0.1-ha plot (Wilson 2014; A. Mariscal, unpublished
data). These forests were extensively cleared in the
1970s and 1980s to produce timber and charcoal, and
for pastures and agriculture (Sarmiento 2002, Zorrilla
2010). Currently, less than 10% of the original low- to
mid-elevation cloud forest cover remains (Sarmiento
1995a, b, Zorrilla 2010) and land is primarily used for
cattle pastures and subsistence and small-scale, market-
oriented agriculture (Kocian et al. 2011; C. Zorrilla,
personal communication; S. Wilson, unpublished data).
Over 80% of residents are mestizo (of mixed indigenous
and Spanish decent), with minority populations of
indigenous Otovaleos and Afro-Ecuadorians.
Individual communities tend to be of mixed ethnicities
(DAmico 2010, Kocian et al. 2011; S. Wilson, unpub-
lished data). Ninety percent of the population falls
below the national poverty line (Kocian et al. 2011).
In the past two decades, international and Ecuadorian
environmental NGOs have worked to create a forest
conservation corridor in the northwest Ecuadorian
Andes, an effort that includes several community-based
restoration projects. In 2001, the local NGO Defensa
y Conservacion Ecologica de Intag (DECOIN) raised
international funding to help communities purchase,
reforest, and protect land in their watersheds (NGO
information available online).5 Our study takes place
in five of these protected watershed reserves where
forests were restored by planting trees. These watershed
reserves are an ideal study site because their land-use
history is typical of many farmed Andean landscapes:
they were cleared 3040 years prior to reforestation,
burned, cropped for a few years, and subsequently
planted with exotic grasses (usually Setaria sphacelata
Schumach., locally called pasto cebolla) and then used
as pasture (Table 2). Reserves were also restored using
local tools and methods that are easily adopted by
communities, and commonly used in Latin America
(Gandolfi et al. 2007; see footnote 4). Technicians
trained residents to collect and propagate seeds from
native trees in nearby forests, and to plant and main-
tain these seedlings in combination with some nonnative
species. Seedlings were planted 2.5 m apart, and people
planted a total of 50 species with between 12 and 26
in each reserve. All reserves were managed similarly.
Community members cleared pasture grass around
5 www.decoin.org
Ecological Applications
Vol. 26, No. 1
seedlings by hand every 34 months, and prohibited
grazing animals, harvesting wood for sale, clearing,
and burning within reserves. In each reserve, there
are planted areas and areas that, because of limited
funds, were not planted but left to regenerate naturally,
creating a control site for each restored area (Table 1).
Forest plots
We interviewed local landholders to find areas replanted
with trees (planted forest, n = 5) and areas allowed to
regenerate naturally (unplanted forest, n = 5) with similar
land-use histories within each reserve. The land use sur-
rounding each is also similar; reserves are located in a
matrix of mainly agricultural land uses, with small (<1
ha) patches of secondary forest present within 500 m.
We also sampled trees in primary forests (primary for-
est, n = 5), located in other reserves in the region (Fig. 1).
Primary is here defined as forests that have not been
cleared in living memory (80 years), are relatively intact,
but may have sustained light selective logging. Few
relatively undisturbed forests remain at this elevation in
Intag; the plots we sampled represent the best examples
of older forests at this elevation in the region (A. Almeida,
personal communication; C. Zorrilla, personal communica-
tion; J. deCoux, personal communication). All sampled
plots in primary and secondary forests were in the same
elevation range (19002250 masl).
At each of the 15 sites, we ran four 50 5 m
transects following slope contours (total 60 transects,
0.1 ha/site). We divided each transect into five 10
5 m plots. In each plot, we counted, identified, and
measured trees >2.5 cm diameter at breast height (dbh;
1.4 m above ground), woody saplings (12.5 cm dbh),
and seedlings (>0.5 m height, <1 cm dbh). Trees were
identified in the field by Ecuadorian botanists, and
we took replicate voucher samples of unknown species
to the Museo Ecuatoriano de Ciencias Naturales
Herbario Nacional del Ecuador (QCNE) in Quito,
Ecuador for identification. A local guide provided us
with common names when possible. Plants were counted
as separate individuals if the stem of the plant was
not connected at or just below the soil surface (Chazdon
et al. 1998). At two random locations on each transect,
we recorded slope, aspect, canopy density, and percent
ground cover in two 1-m2 plots.
Dispersal mechanisms and plant characteristics
Dispersal mechanisms for common species (comprising
at least 0.2% of the total number of stems in our data
set) were determined by botanists at QCNE. These data
were verified and supplemented with other published
sources. If more than one dispersal mechanism was pos-
sible, we selected animal over wind or gravity, and relied
on information from local botanists over published sources,
as dispersal mechanisms can be different in different
locales (Chazdon et al. 2003). QCNE botanists also
January 2016 COMMUNITY-BASED CLOUD FOREST RESTORATION
8 1
0 3 6 km
FIG. 1. Map of study sites in Intag, Ecuador. Numbers 15 are watershed reserves, each of which contains an area of planted
forest and an area of unplanted forest. Numbers 610 are primary forest reserves.
recorded the type of forest where each species is typically
found (secondary or primary) based on herbarium infor-
mation and their extensive field experience.
Soils
At each site, we took 10 soil samples from the top
10 cm of soil at two to three randomly located spots
on each transect. We created one composite soil sample
per site, stored them in a conventional refrigerator,
and delivered them to a soil laboratory within 15 d
(Estacion Experimental Santa Catalina, Instituto
Nacional Autonomo de Investigaciones Agropecuarias,
Cutuglagua, Meja, Pichincha, Ecuador). Soils were
analyzed for macronutrients (nitrogen [NH4], phospho-
rus [Olsen extraction method], potassium), organic
matter content, and texture (Santa Catalina information
available online).6 Four bulk density samples were taken
using a cylindrical sampler 10 cm in diameter from
the top 10 cm of soil at each site (one per transect)
and weighed while wet and then periodically after
being sun-dried for several weeks until the dry mass
had stabilized.
Species identified as useful by people
In focus groups (n = 6), interviews with local experts
(n = 10), and household surveys (n = 120), we asked
participants to identify which trees people use, or would
use if available, and for what purpose (e.g., medicine,
firewood, live fences, construction). We also asked
people what forest products they currently harvest and
6 http://www.iniap.gob.ec/sitio/index.php?option=com_cont
ent&view=article&id=36&Itemid=15
207
N
4
2
5
3
6
9
7
ECUADOR
10
for what purpose. Finally, we conducted oral histories
(n = 16) with older community residents about changes
in forest cover and use over time.
Analysis
We made rarefaction curves using the program
EstimateS (version 8.2.0; Colwell 2009) to compare
species richness in primary, planted, and unplanted
forests. We resampled the data 100 times to create
smooth species accumulation curves (showing the addi-
tional number of species found as more individuals
are sampled) with 95% confidence intervals that can
be used to determine whether two sites contain sig-
nificantly different species counts for a given number
of individuals (Chazdon et al. 1998, Colwell 2009,
Colwell et al. 2012). We compared the species richness
and species density (cumulative number of species per
area sampled) between planted and unplanted forests
in each reserve separately, and with all sites of each
forest type pooled. We then rarefied species richness
to a common number of stems at each site and com-
pared these values across sites using analysis of vari-
ance (ANOVA) tests (SPSS version 20.0, IBM, Armonk,
New York, USA). In planted forests, we assessed the
correlation between the number of species initially
planted and current species richness.
We used nonmetric multidimensional scaling (NMDS)
ordination to assess similarity in the species composi-
tion (i.e., both the types and relative abundances of
species at each site) of trees in primary, planted, and
unplanted forests. The Chao-Jaccard similarity estima-
tor was used as a distance measure to help correct
for the undersampling bias often present in extremely
biodiverse ecosystems (Chao et al. 2005, Norden et al.
208 S. J. WILSON AND J. M. RHEMTULLA
A sary, variables used in ANOVAs were square-root or
Planted secondary forest excluding planted trees
Planted secondary forest
Unplanted secondary forest
Primary forest
B Bonferroni post hoc, P = 0.005, P = 0.013), which
c both tended to be higher in primary forest soils than
FIG. 2. Species richness and species density. (A) Species
richness in individual sites: primary (green), planted secondary
(light blue; PL), and unplanted secondary (red; UP). Numbers
(110) represent different sites. (B) Species richness pooled in
primary (green), planted secondary (light blue), planted
secondary with the planted trees excluded (dark blue), and
unplanted secondary (red) forest. (C) Species density in
pooled primary, planted, and unplanted forest for a total 1-ha
area. Note that species density allows comparison of values at
a given sampling effort and allows us to assess if the sampling
effort was sufficient to compare this value across sites. In (B)
and (C) dotted lines indicate 95% confidence intervals based
on 100 randomized runs. Note the different values on the x-
axes.
2009). To compare species compositions across forest
types and size classes, we used a multi-response per-
mutation procedure (MRPP). To examine successional
patterns, we separately ordinated seedlings, saplings,
and trees at each site (Norden et al. 2009). In addi-
tion, to determine if tree species that regenerate under
Ecological Applications
Vol. 26, No. 1
planted trees differ from those that regenerate on
unplanted pasture, we removed all the trees that had
been planted from the data set and re-ran the diversity
and species composition analyses.
We included soil and ground-cover data as explana-
tory variables in the multivariate analyses, and statisti-
cally compared soil data between forest types using
ANOVA. We also compared the absolute and relative
abundance of anthropogenically useful species in the
three forest types using nested ANOVAs. When neces-
natural log transformed before analysis to meet assump-
tions of normality.
RESULTS
Soils
Planted and unplanted sites had similar chemical
and physical soil properties to one another, but primary
forest soils were distinct. Soil bulk density was lower
in primary forests than in either planted or unplanted
secondary forests (ANOVA, F2,14 = 9.09, P = 0.004,
were not significantly different from one another
(Bonferroni post hoc, P = 0.861; all results significant
at P < 0.05). Primary forest soils also contained more
organic matter than either planted or unplanted sec-
ondary forest (ANOVA, F2,14 = 7.39, P = 0.008,
Bonferroni post hoc, P = 0.013, P = 0.027). Primary
forest soils were generally classified as organic, and
secondary forest soils as mineral (Appendix S1; see
footnote 3). Nitrogen (NH4; ANOVA, F2,14 = 1.00, P
= 0. 400) and phosphorus were similar across forest
types (ANOVA, F2,14 = 1.52, P = 0. 250), although
in either planted or unplanted secondary soils.
Potassium, however, was higher in planted forest soils
than in primary forest soils (ANOVA, F2,14 = 4.23,
P = 0.040, Bonferroni post hoc, P = 0.010; Table 3).
Although it should be noted that the small sample
size may have masked variation, overall there was little
evidence of structural (i.e., bulk density) or chemical
soil recovery in planted sites as compared to unplanted
ones.
Tree communities
We identified a total of 6936 individual trees and
416 woody plant species in all 15 sites. The majority
of species were trees (345) and woody shrubs (55),
with some tree fern species (11), and woody species
for which the form was unknown (5). For simplicity,
from here on we refer to all species as trees (Norden
et al. 2009). Unplanted, planted, and primary forests
had a total of 58 (mean per site 15.4 3.8; all means
presented standard deviation), 129 (44.2 17.4),
and 300 (88.2 24.1) species, respectively. In total,
January 2016 COMMUNITY-BASED CLOUD FOREST RESTORATION 209

6SE
3TR 6SA

1SE
3SE
3TR 4SA
3SA 5SE 6TR
1SA
3SA Native grass
2SE Bare ground
1TR 5TR 9TR Wind-dispersed
9SE 8TR A2
3SE 4SA 10SE
2SA 4SE Native herbs
2TR
Moss Animal
Wind 1SE Bracken fern 4TR 8SE 7SE 7SA
4SE
Exotic grass
5SE 8SA 10SA
1TR 5TR 5SA 10TR

2TR 2SA 4SA 5TR

4TR
Axis 2

9SA
2SE
1SA A1
A3 Animal-dispersed

6SE
3TR
6SA
1SA

3TR 3SE 3SA 5SA 6TR

1SE
3SE 5SE

3SA 2SA 9SE

4SE 9TR
1TR 4SA 8TR B2 Wind-dispersed
10SE
1SA
2SE 2TR 2TR 8SE
5SA 10TR
4SA
5SE 8SA
1TR 10SA
5TR 7SA

4SE Forest Type 7SE

Primary 9SA
2SE
2TR 1SE Planted secondary
4TR Unplanted secondary 7TR
2SA
B1 B3 Animal-dispersed
Early sucessional forests Later sucessional forests
Wind dispersed and pioneer species Axis 1 Animal dispersed and primary forest species

FIG. 3. Nonmetric multidimensional scaling ordination for primary (green), planted (turquoise), and unplanted (orange)
forests. Numbers 15 are secondary forest sites, 610 are primary forest sites. At each site, trees were divided by size class (see
Methods: Forest plots for sizing criteria). Circles are seedlings (SE), squares are saplings (SA), and triangles are trees (TR). Part
1 is an ordination of all trees in planted (Pl), unplanted (Up), and primary (Pr) forests, and arrows represent ground cover
characteristics that explained more than 25% of the variation in the data (r2 > 0.25). Part 2 is an ordination excluding planted
trees (only trees naturally regenerating in planted forests are included). Parts A1 and B1 show that wind- and animal-dispersed
species clearly separate along the x axis, A2 and B2 show wind-dispersed species and their relative influence in different sites
and size classes, as indicated by the size of the symbol for each, and A3 and B3 show the same for animal-dispersed species. Both
are two-dimensional models that explain 52% and 50% of the variation in the data, with stress levels of 14.32 and 13.43,
respectively.

people planted 51 different trees species (1233 per species-rich planted forest (Fig. 2A, B). Species rich-
site; Table 1). ness was higher in planted secondary forest than
unplanted secondary forests both when sites were
compared individually (except for site 3; Fig. 2A)
Stem density and species richness
and pooled (Fig. 2B). Following the same pattern,
Mean stem density was not significantly different species richness rarefied to a common number of
between primary forest (6590 stems/ha 1683) and stems was significantly different across forest types,
planted forest (5690 3130), but both were signifi- with primary forests having the highest (28.0 2.7),
cantly higher than unplanted forest (1474 1207; planted forests the next (18.2 3.8), and unplanted
ANOVA, F2,14 = 10.91, P = 0.020, Bonferroni post forests having the least (12.6 2.3) number of spe-
hoc, P = 0.009, P = 0.737). Primary forests were cies (ANOVA, F2,14 = 33.21, P = 0.000; Bonferroni
more species rich than either type of secondary for- post hoc, all P < 0.040). Because the absolute stem
est, although species richness in the least rich primary density of trees was low in unplanted forest, the
forest was comparable to that of the most difference in species density between unplanted (17.6
210 S. J. WILSON AND J. M. RHEMTULLA
species/0.1 ha 7.3) and planted forests (44.2
17.4) was even greater than the difference in species
richness. Primary forests had the greatest species
density (88.0 24.2; Fig. 2C; ANOVA, F2,14 = 20.21,
P = 0.000; Bonferroni post hoc, all P < 0.020).
There was no detectable relationship between the
number of species planted initially and species rich-
ness (r = 0.15, P = 0.809).
Species composition
In the ordination, sample points (which each rep-
resent a size class [seedlings, saplings, trees] at a given
site) separated by forest type on axis 1 with unplanted
forests on the left, planted in the left-center, and pri-
mary on the right (Fig. 3:A1). Forest type was a
highly significant grouping variable (MRPP, t = 4.40,
P = 0.001). Communities of seedlings, saplings, and
trees at the same site tended to be similar: points
grouped by site, not size class (MRPP of size class,
t = 2.50, P = 1.000). Planted and unplanted sites
tended to group by geographical location and/or age:
sites 5 and 4, two geographically close, older sites
grouped together, as did the younger and also geo-
graphically close sites 2 and 3 (Figs. 1, 3:A1, Table 1).
The primary forests tended to group secondarily by
their geographical location in the study valley (Fig. 1);
primary forests sites in the northeast (sites 6 and 9)
were in the upper part of the ordination, and the
three southwestern sites (sites 7, 8, 10) in the lower
part. Plots in planted forests tended to group more
tightly in ordination space than unplanted and primary
forests, which implies that species composition in
planted forests was more homogenous (Fig. 3:A1).
Common species (0.2% of overall abundance)
accounted for 83% of all trees, including 73% of trees
in primary, 91.2% in unplanted, and 90.9% in planted
forests. Most species were dispersed by animals (68%;
primarily birds) or wind (27%). Wind-dispersed species
characterized unplanted forests and showed a strong
negative correlation with axis 1 (r = 0.87, Fig. 3:A2);
animal-dispersed species, with primary forest, showed
a strong positive correlation with axis 1 (r = 0.90,
Fig. 3:A3). All planted sites contained at least some
animal-dispersed species (mean 50.0%, SD 28, range
21.770.1%). Some older unplanted sites also contained
animal-dispersed species, but younger unplanted sites
contained very few (mean 24.2%, SD 20.0, range
3.143.9%). The majority of species in primary forests
were animal dispersed (87.1%, SD 2.6, range 85.8
91.8%). In primary forests, we found 27 species that
were on the UN red list for species of high conserva-
tion priority. These species were largely absent from
secondary forests; only two were present in planted
forests, and none in unplanted.
Secondary forests, especially unplanted sites, were
strongly associated with nonnative ground covers such
as the exotic pasture grass Setaria sphacelata (Schumach;
Ecological Applications
Vol. 26, No. 1
Abundance (species per 0.1 ha)
300
Multipurpose
Timber
250 Firewood
200
150
100
50
Unplanted Planted Primary
FIG. 4. The abundance of species that people use in unplanted,
planted, and primary forests, divided by use. Error bars show
standard deviation.
Poaceae) and bracken fern Pteridium aquilinum (L;
Dennstaedtiaceae), an invasive pan-global species
(Schneider 2004). Planted secondary forests tended to
have more bare soil and native grasses, and primary
forests tended to have more moss, leaf litter, and
native, non-grass herbaceous plants (Fig. 3:A1).
To further compare the secondary forest types, we
ran a NMDS without the primary forest stands
(Appendix S2). This made it possible to distinguish
which species were associated with planted and
unplanted secondary forests. Older, planted sites were
associated with species that were mostly animal dis-
persed from a range of families: Piper aduncum (L.;
Piperaceae), Urera caracasana (Jacq.; Urticaceae),
Palicourea demissa (Sandl.; Rubiaceae), Carica pube-
scens (Lenne and C. Koch; Caricaceae), Inga sp.
(Fabaceae), and Croton mutisianus (Kunth.;
Euphorbiaceae, wind-dispersed). In contrast, younger,
unplanted forests were mainly characterized by shrubby
or small tree species with lightweight wood and wind-
dispersed seeds in the Asteracaeae family, primarily
Baccharis latifolia (Ruiz and Pav.) and B. trinervis
(Lam.; Asteraceae). Results otherwise were similar to
the NMDS with primary forest sites.
Tree recruitment in unplanted and planted secondary
forest (excluding planted trees)
To examine how planting trees affects tree recruit-
ment, we compared patterns in naturally regenerating
trees in planted and unplanted sites by reanalyzing
the data after removing trees planted by people.
Twice the number of trees, seedlings, and saplings
were naturally regenerating in planted sites (1580
individuals) than in unplanted sites (738). Both the
density and the diversity of naturally regenerating
trees were higher in planted than in unplanted sites
(Fig. 2B). The species composition of naturally
January 2016 COMMUNITY-BASED CLOUD FOREST RESTORATION
regenerating stems in planted forest was significantly
different from both unplanted and primary forests
(MRPP, t = 3.62, P = 0.005), but it was more
similar to unplanted forests than when the planted
species were included (Fig. 3:A1, B1). Three key pat-
terns emerged: (1) seedlings and saplings in planted
forest were most similar to primary forests, and dis-
similar to unplanted forests; (2) trees in planted forest
grouped with seedlings and saplings in unplanted
forest; and (3) trees in unplanted forest were the
least similar to primary forests. Size class was not
significant (MRPP, t = 2.23, P = 1.000).
Planted forests had a much greater proportion of
animal-dispersed trees naturally regenerating than did
unplanted forests (Fig. 3:B2, B3), and the understories
of both planted and unplanted secondary forests con-
tained relatively more animal-dispersed species than their
overstories. Of all naturally regenerating individuals in
planted areas, almost half (43.3%) were animal-dispersed
species (50.9% were wind dispersed) compared to only
23.7% in unplanted areas (where 75.6% were wind dis-
persed). Across size classes, relatively few naturally
regenerating trees in planted forests were animal dispersed
(23.4%), while a much higher proportion (47.6%) of
the regenerating seedlings and saplings (the smaller size
classes) were animal dispersed. In comparison, in
unplanted forests, 8% of the trees and 27.7% of the
seedlings and saplings were animal dispersed. Overall,
this ordination suggests that both types of secondary
forests are undergoing succession, but that planted forests
are accumulating more animal-dispersed species earlier,
and at a faster rate, than unplanted forests.
Species people plant and use
People used trees mainly for firewood and timber
(44% and 40% of the useful species, respectively). Other
species were used for fertilizer/nitrogen fixation (20%),
food (16%), livestock fodder (8%), and medicine (8%;
note that seven species were multiuse). Planted forests
contained more useful species than both primary and
unplanted forests (ANOVA, F2,13 = 11.72, P = 0.020;
Bonferroni post hoc, P = 0.020, 0.001; Fig. 4). The
relative abundance of useful species also differed among
forest types (ANOVA, F2,13 = 6.63, P = 0.012) and
was lower in primary forests (3% 2%) than in planted
secondary forests (Bonferroni post hoc, P = 0.011),
but similar in unplanted (21% 13%) and planted
(40% 18%) secondary forests (Bonferroni post hoc,
P = 0.360). Planted forests contained proportionately
more species that were used for timber, while most
useful species in unplanted forests were used exclusively
for firewood (2 = 264, df = 1, P < 0.000).
DISCUSSION
The results suggest that putting restoration projects
under community management can fulfill a fundamental
211
goal of ecological forest restoration; to assist the
recovery of an ecosystem that has been degraded,
damaged or destroyed (Clewell et al. 2004:3). In this
study, community-managed tree-planting efforts quickly
increased the species richness of secondary forests and
also jump-started ecological succession, increasing the
number of tree species, and in particular animal-
dispersed species, in the understory. Early forest recov-
ery was greatly enhanced by local planting efforts.
Planting trees accelerates forest recovery over
spontaneous regeneration
Restoration efforts clearly increased forest recovery
on pastures in Intag over the time frame studied; after
four to seven years, forest recovery without tree plant-
ing was minimal, with low species richness. Unplanted
forests consisted of a few wind-dispersed, shrubby tree
species with soft wood; species that commonly colonize,
and sometimes dominate, Andean pastures (Zahawi and
Augspurger 1999, Posada et al. 2000, Amzquita et al.
2004). These species are functionally and structurally
distinct from most species found in primary cloud for-
est (Chazdon et al. 2003). Exotic pasture grasses and
ferns were abundant in the understory of unplanted
sites which, along with a lack of nearby seed sources
and animal dispersers, can prevent forests from regen-
erating (Holl et al. 2000, Brockerhoff et al. 2008,
Griscom et al. 2009, Aide et al. 2010, Ortega-Pieck
et al. 2011) and lead to a state of arrested succession
(Zahawi and Augspurger 1999, Posada et al. 2000).
In contrast to unplanted forests, young planted forests
had more trees, more species of trees, and more types
of tree species. Planting trees helped forests recover
quickly both by (1) directly reestablishing trees, and
(2) improving site conditions to facilitate the recruitment
of other tree species in the understory. People planted
a variety of trees, selecting species with specific char-
acteristics such as durable, fast-growing wood (timber
trees), nitrogen-fixing symbionts, and fleshy, edible fruits.
This planting created species-rich forests composed of
trees with different functional traits from those in
unplanted forests. Most naturally regenerating trees were
new recruits from outside the planted site; the majority
of them were different species from the planted trees.
Young forests undergoing succession should, in theory,
begin to accumulate the same species of seedlings and
saplings as adult tree communities in primary forest
(Terborgh et al. 1996, Norden et al. 2009), a pattern
which we observed in planted sites, especially older
ones. Although communities of naturally regenerating
trees were similar to unplanted forests, the seedlings
and saplings in planted forests were distinct from
unplanted forests, and more similar to primary forests.
Thus, only four to seven years after planting, planted
forests had accumulated more and different tree species
than unplanted forests, and their species composition
was relatively, and increasingly, more similar to primary
212 S. J. WILSON AND J. M. RHEMTULLA Ecological Applications
Vol. 26, No. 1

forests. Because naturally regenerating secondary forests
can retain distinct communities of trees from primary
forests for decades, and initial differences in species
compositions can persist over long time periods (Dent
and Wright 2009, Klanderud et al. 2010, Chai and
Tanner 2011, Martin et al. 2013), community-based tree
planting efforts have the potential to set regenerating
forests on a different successional pathway than natural
regeneration. Although our study examines young forests,
planting trees to restore forests could have longer-term
impacts on forest species composition in this region.
It seems likely that planting trees jump-started suc-
cession in part by attracting seed dispersers. A lack
of seed dispersers is one of the largest barriers to
tropical forest recovery, especially in landscapes where
primary forest cover is low (Holl et al. 2000, Chazdon
et al. 2003, Vieira et al. 2009, Aide et al. 2010), as
was the case at our sites (S. Wilson, unpublished data).
Because people planted many animal-dispersed species,
they may have created forests more capable of pro-
viding birds and mammals with habitat and food (Holl
1998, de la Pea-Domene et al. 2013). In tropical
forests, most pioneer species are wind-dispersed, but
most tree species in primary tropical forests are animal-
dispersed (at our sites, a mean of 89%). Ensuring
that animal-dispersed species regenerate is thus a pri-
ority for restoration (Holl 1999, Holl et al. 2000,
Chazdon et al. 2003, Aide et al. 2010). Experimental
plot-based studies agree with our results, showing that
tree planting can accelerate early successional processes,
particularly when people plant animal-dispersed species
(de la Pea-Domene et al. 2013, Jacob 2014).
In addition to increasing dispersal, planting and
maintaining trees changed site conditions. Poor site
conditions can prevent many tree species from estab-
lishing, even if they arrive (Holl 1998, Holl et al.

TABLE 2. Species that people use as defined by community members.

Species Local name Uses Forest type Restoration notes

Alnus accuminata H. B. and K. A. aliso soft timber (fences, fur- PL nonnative, fast growth, easy
niture, framing), sil- to propagate, seeds
vopastoral, soil harvested from farms
Amasonia sp. atambo firewood PL, UP very soft wood
Baccharis nitida Ruiz and Pav chilca firewood PL, UP very soft wood
Baccharis latifolia Pers. chilca firewood PL, UP very soft wood
Baccharis sp. pichulan firewood PL, UP very soft wood
Brunellia cf. acostae Cuatrec. fresno timber P, PL produces seeds often, easy to
propagate
Calliandra pittieri Standil. tura firewood, fertilizer PL
(leaves), silvopasture,
soil
Casearia sp. pilche, pilche timber
blanco
Cedrela odorata L. cedro hard timber (furniture, PL seeds brought from outside
houses) community
Cinchona sp. cascarillo medicinal P, PL
Clusia alata Planch. and Triana guandera firewood P, PL food for birds
Clusia crenata Cuatrec. guandera firewood P, PL food for birds
Ocotea glaucosericea Rowher laurel timber PL food for birds
Croton floccosus B.A.Sm. drago timber, medicinal PL
Delostoma integrifolium D. Don dialoman firewood, silvopastoral P, PL, UP
Erythrina edulis Micheli poroton firewood, food for PL
guinea pigs (Cavia
porcellus), fences, ed-
ible seeds
Guaiacum sp. guayacan hard timber (houses) PL
Hyeronima scabrida Tul. motilon timber PL
Inga densiflora Benth. guaba food, soil P, PL
Juglans neotropica Diels. nogal timber PL
Leucaena sp. leucaena feed for guinea pigs, NA nonnative, used on farms
soil live fences
Myrcianthes cf. orthostemon O. Berg cungla firewood PL, UP
Nectandra purpurea Ruiz and Pav canelo hard timber P
Persea americana Mill agucate food, firewood P, PL, UP food for birds
Saurauia brachybotrys Buscal. moco firewood, edible fruit P, PL, UP

Notes: Uses are those stated by the community members. Forest type refers to the forest in which each species is found, determined
through plant surveys; abbreviations defined in Table 1.
January 2016 COMMUNITY-BASED CLOUD FOREST RESTORATION
2000, Chazdon et al. 2003, Aide et al. 2010, Ortega-
Pieck et al. 2011). Exotic pasture grasses (Setaria
sphacelata) and invasive bracken ferns (Pteridium
aquilinum) dominated the understories of unplanted
forests, but were less abundant in the understories of
planted forests (Holl et al. 2000). The increase in
canopy cover at planted sites could have, in part,
shaded out S. sphacelata, allowing primary forest tree
species to regenerate, as has been observed at other
sites (Rhoads, Eckert, and Coleman 1998). In addi-
tion, although some other studies have found that
removing grass decreased or did not affect tree survival
(Holl 1999), according to local people in Intag, remov-
ing exotic grass appeared to increase tree survival in
the initial phases of the project. Once trees were
established and people stopped weeding (two to three
years after planting) understories of planted forests
still contained fewer exotic plants.
As has been found in past studies over a similar
time frame, planting trees did not affect soil bulk
density, organic matter content, or macronutrient con-
tent in the short term (Holl and Zahawi 2014). Soil
in secondary forests was still nearly twice as dense as
in primary forests. Although compacted soils can pre-
vent trees from establishing and thriving, and are a
major barrier to cloud forest regeneration in abandoned
pastures (Kozlowski 1999, Pedraza and Williams-Linera
2003, Aide et al. 2010), in this case planting trees
facilitated tree establishment in the short term without
noticeably improving soils.
Planted forests have high use value
Planted forests in this study contained proportionately
more species that people use, and more multipurpose
and timber species, than unplanted regenerating forest
and primary forests. Community members preferred
to plant tree species: (1) that can be easily propagated
and transplanted; (2) that grow quickly; (3) that have
cultural significance, such as the wax palm, Ceroxylon
ventricosum (Burret), from which people harvest fronds
to celebrate the religious holiday Palm Sunday; and
(4) that have local use value (Table 2). People in
Intag reported that although in the past they used
several forest species, they learned of and adopted
uses for several additional timber, food, fodder, and
medicinal species during restoration. Planted forests
thus contain more useful species both because people
preferred to plant species that they already used, and
because the NGO identified uses for other trees, includ-
ing technical information on how native trees could
enhance farming systems and provide timber. This type
of knowledge sharing is exactly what many community-
based programs aim to foster (Diemont et al. 2011,
Uprety et al. 2012). Smallholders around the world
plant trees as part of farming and forestry systems
(Diemont et al. 2011, Hoch et al. 2012). Tapping into
this local knowledge can expand the pool of species
213
that can be found, propagated, and for which growth
requirements are known (Surez et al. 2012).
Project managers and communities used both social
and ecological criteria to select tree species for plant-
ing. Two exotics were planted for economic reasons:
citrus trees (Citrus spp.), which produce a highly
consumed, locally marketable food; and alders (Alnus
nepalensis), fast-growing, nitrogen-fixing trees used for
lightweight timber, firewood, fence posts, and improv-
ing soil fertility. These species were initially included
to motivate people to participate: they had prior expe-
rience planting them (citrus), knew that they produced
relatively quickly (alders), and were known to survive
in harsh environments. However, the NGO also required
people to grow and maintain a high ratio of native
to exotic trees; an example of selecting species for
ecological reasons (Zorrilla 2010). Many species were
planted for both ecological and social reasons; native
timber trees were selected over exotic ones, for example.
Our results show that restoration projects are well
suited to, and can benefit greatly from, combining
local or traditional ecological and scientific knowledge
to select and propagate species (Kappelle et al. 2000,
Kirby and Potvin 2007, De Koning et al. 2011, Diemont
et al. 2011, Uprety et al. 2012).
The future of community-planted forests: restoring
forests, or creating novel ones?
In our study, community-based tree-planting projects
are working to restore species-rich forests that contain
more animal-dispersed species and are more structur-
ally similar to primary forests than unplanted forests
are. But, these forests are still not guaranteed to become
the forests seen in the past: the species composition
of planted and primary forests remained distinct in
ways that may persist over time.
Our primary forest sites contained 27 tree species
on the IUCN Red List; only two of these were found
in our planted sites (intentionally planted because of
their conservation status), and none in unplanted forest
(IUCN Red List available online).7 This finding may
be symptomatic of another pattern that we observed:
that planted forests in different sites were more homog-
enous than either unplanted or primary forests.
Homogenization has been observed in secondary forests
that have spontaneously regenerated (McKinney and
Lockwood 1999, Holl 2002, Lugo and Helmer 2004,
Rhemtulla et al. 2007). Rare or endemic species may
thus be lost, which is of particular concern in tropical
montane cloud forests characterized by high numbers
of endemic species and high species turnover on a
landscape scale (Chai and Tanner 2011, Jost 2013,
Martin et al. 2013). Indeed, our primary forest sites
exhibited distinct species compositions, with communi-
ties of seedlings, saplings, and trees grouping by site,
7 http://www.iucnredlist.org/
214 S. J. WILSON AND J. M. RHEMTULLA
rather than by size class, as has been observed in
some lowland tropical forests where the composition
of tree communities varies less in space (Condit et al.
2002, Norden et al. 2009).
People in our study planted a total of 50 species
(1226 per site), a relatively large number compared
to other projects that have typically planted five
to 25 (Fang and Peng 1997, Ruiz-Jaen and Aide
2005, Leopold and Salazar 2008, de la Pea-Domene
et al. 2013; but see Knowles and Parrotta 1995,
Parrotta et al. 1997). But this is still a small subset
of the more than 300 species found in the primary
forests in this study. We currently lack information
about many cloud forest tree species (physiology,
growth requirements, reproductive cycles, etc.) that
would allow us to propagate them. It is thus inevi-
table that a selection process in which useful species,
or other species found to work well in restoration,
will be widely used and propagated across the land-
scape (Knowles and Parrotta 1995). This species
bias could mean that rather than recreating primary
forests, restoration may create novel forests (Lugo
2009) that can persist for decades, if not centuries
(Lamb et al. 2005, Dent and Wright 2009, Klanderud
et al. 2010, Chai and Tanner 2011, Martin et al.
2013).
The potential to create novel forests through tree
planting is not unique to community-based forest
restoration. Currently, any tropical restoration project
is unlikely to have the resources to replant the entire
suite of tropical forest species. The use of exotic
timber plantations to jump-start forest recovery, for
example, is a contentious issue: although trees can
often regenerate in plantation canopies, we still do
not know the long-term effects of introducing these
exotics into native forests (Lugo 1997, 2009). How
acceptable these changes in species composition are
will vary with the social and ecological goals of a
given project (Higgs 2003, Holl 2011). Community-
based projects have an advantage here: at least local
people can tailor these inevitable differences in biotic
communities to meet their needs. Long-term studies
are needed in all of these reforestation projects to
determine if succession will proceed in the direction
TABLE 3. Physical and chemical soil properties in planted, unplanted, and primary forests.
Soil variable Primary
pH 5.05 0.97
Bulk density (g/cm3) 0.47 0.12
Soil organic matter (%) 25.6 10.7
NH4 (mg/kg) 59.9 38.7
P (Olson modified; mg/kg) 12.8 7.0
K (Olson modified; mg/kg) 210 110
Notes: Values for primary, planted, and unplanted are given as mean SD. Soil matter is shown as a percentage of mass. For F,
differences are significant at P < 0.05; n = five sites in each treatment.
Ecological Applications
Vol. 26, No. 1
of a primary forest, or if planting a higher abun-
dance of particular species will alter the composition
of secondary forests permanently, resulting in novel
forests.
Implications for restoration management and policy
Policy should encourage projects to propagate and
plant species that fulfill dual ecological and social
goals. In particular, project managers could promote
locally used species that also attract seed dispersers,
such as food-producing trees with large, fleshy fruits,
or hardwood timber species, which are also often ani-
mal dispersed.
Making local communities central to planning, main-
taining, and protecting planted forests can make pro-
jects more relevant to local people and create local
stewardship (Agrawal and Angelsen 2009, Wilson
2013). We found that the time and effort that local
people invested in reforestation; 75 households work-
ing up to 50 days/year for one to three years (S.
Wilson, unpublished data), also made people more
intent on protecting these sites. When asked if they
planned to clear planted forests, people unanimously
replied no, often explaining that they were planting
trees to create a permanent forest. Local people com-
monly referred to planted forests as buen bosque (good
forest), while naturally regenerating areas were not
even called forest, but chaparro; scrubby brush com-
monly cleared for agriculture (some farmers mentioned
plans to cut and replace chaparro with planted
forest).
But practitioners should not assume that local people
make choices that are optimal for specific environmental
conservation goals (Robbins 2001). If threatened or endan-
gered species are to be conserved, for example, this must
be an explicit project goal. Projects could also provide
long-term funding for enrichment planting with rare or
threatened forest species under established canopies.
Finally, as we have demonstrated here, studying
community restoration sites can provide valuable
insights on the ecological and social outcomes of res-
toration; information that can improve future restora-
tion projects. In addition to experimental plots studies,
Planted Unplanted F P
6.02 0.16 5.85 0.22 0.44 0.650
0.81 0.22 0.78 0.17 9.09 0.004
12.1 3.0 13.7 6.4 7.40 0.008
33.6 3.2 39.6 19.1 1.00 0.400
7.44 3.6 9.82 3.8 1.54 0.250
488 226 324 129 4.23 0.041
January 2016 COMMUNITY-BASED CLOUD FOREST RESTORATION
researchers can study the field-based experiments inher-
ent in so many restoration projects. Projects provide
opportunities for researchers and local people to collect
data examining questions at several stages: (1) how
to propagate different local species in nurseries, (2)
which local species grow best under what conditions,
and (3) how plant communities reassemble once spe-
cies are planted in the field (Menninger and Palmer
2006, Lovett et al. 2007). Indeed, in the course of
this research, the principal investigator encountered
local research and monitoring projects in more than
a dozen communities, which, were they to be published,
would provide valuable data to scholars and practi-
tioners alike from previously unstudied tropical forest
regions.
ACKNOWLEDGMENTS
We wish to thank the many people of Intag who partici-
pated in this project. Thank you to botanists Miguel Angel
Chinchero, Jenny Elizabeth Correa, Gabriela Cruz, and Carlos
Morales for their help in the field and lab; to Miriam Harder,
Silvana Bolanos, Alonzo Andrengo, and Carmen Navarette
for their assistance with interviews; to Carlos Zorilla of
DECOIN and to Ana Mariscal of Fundacion Cambugan for
their help with fieldwork logistics and information; to Oliver
Coomes, Sylvia Wood, Ignacia Holmes, and Aerin Jacob for
their insightful comments on results and earlier versions of
this article; and to Jake Brennan for his assistance with soil
sampling and interviews, and constructive comments on this
article. This study was funded by the International
Development and Research Centre (IDRC), the National
Science and Engineering Council of Canada (NSERC), the
Fonds de Recherch Nature et Technologies Quebec (FQRNT),
and the Theo Hills foundation.
LITERATURE CITED
Adams, W. M., R. Aveling, D. Brockington, B. Dickson, J.
Elliott, J. Hutton, and W. Wolmer. 2004. Biodiversity
conservation and the eradication of poverty. Science
306:11461149.
Agrawal, A., and A. Angelsen. 2009. Using community forest
management to achieve REDD+ goals. Pages 201212 in
A. Angelsen, M. Brockhaus, M. Kanninen, E. Sills, W.
D. Sunderlin, and S. Wertz-Kanounnikoff. Realising
REDD+: national strategy and policy options. Center for
International Forestry Research, Bogor, Indonesia.
Agrawal, A., A. Chhatre, and R. Hardin. 2008. Changing
governance of the worlds forests. Science
320:14601462.
Agrawal, A., and C. C. Gibson. 1999. Enchantment and
disenchantment: the role of community in natural resource
conservation. World Development 27:629649.
Aide, T. M., and J. Cavelier. 1994. Barriers to lowland
tropical forest restoration in the Sierra Nevada de Santa
Marta, Colombia. Restoration Ecology 2:219229.
Aide, T. M., M. C. Ruiz-Jaen, and H. R. Grau. 2010. What
is the state of tropical montane cloud forest restoration?
Pages 101109 in L. A. Bruijnzeel, F. N. Scatena, and L.
S. Hamilton, editors. Tropical montane cloud forests: science
for conservation and management. Cambridge University
Press, Cambridge, Massachusetts, USA.
Amzquita, M. C., M. Ibrahim, T. Llanderal, P. Buurman, and
E. Amzquita. 2004. Carbon sequestration in pastures,
215
silvo-pastoral systems and forests in four regions of the Latin
American tropics. Journal of Sustainable Forestry 21:3149.
Baillie, I. C. 1996. Soils of the humid tropics. Pages 256286
in P. W. Richards, editor. The tropical rain forest: an
ecological study. Cambridge University Press, Cambridge,
Massachusetts, USA.
Berry, N., and C. Clunas. 2009. Neblina Reserve Conservation
Project: technical feasibility of a forest carbon project.
Econometrica. Rainforest Concern Report RFCN09001.
Rainforest Concern, Bath, UK.
Bonner, M. T., S. Schmidt, and L. P. Shoo. 2013. A meta-analytical
global comparison of aboveground biomass accumulation between
tropical secondary forests and monoculture plantations. Forest
Ecology and Management 291:7386.
Brockerhoff, E., H. Jactel, J. A. Parrotta, C. Quine, and J.
Sayer. 2008. Biodiversity and planted forests: oxymoron
or opportunity? Biodiversity & Conservation 17:925951.
Chai, S. L., and E. Tanner. 2011. 150-year legacy of land
use on tree species composition in old-secondary forests
of Jamaica. Journal of Ecology 99:113121.
Chao, A., R. L. Chazdon, R. K. Colwell, and T. J. Shen.
2005. A new statistical approach for assessing similarity
of species composition with incidence and abundance data.
Ecology Letters 8:148159.
Charnley, S., and M. R. Poe. 2007. Community forestry in
theory and practice: where are we now? Annual Review
of Anthropology 36:301336.
Chazdon, R. L. 2003. Tropical forest recovery: legacies of
human impact and natural disturbances. Perspectives in
Plant Ecology, Evolution, and Systematics 6:5171.
Chazdon, R. L. 2008. Beyond deforestation: restoring forests
and ecosystem services on degraded lands. Science
320:14581460.
Chazdon, R. L., S. Careaga, C. Webb, and O. Vargas. 2003.
Community and phylogenetic structure of reproductive traits
of woody species in wet tropical forests. Ecological
Monographs 73:331348.
Chazdon, R. L., R. K. Colwell, J. S. Denslow, and M. R.
Guariguata. 1998. Statistical methods for estimating species
richness of woody regeneration in primary and secondary
rain forests of northeastern Costa Rica. Pages 285310 in
F. Dallmeier, and J. A. Comiskey, editors. Forest biodiversity
research, monitoring and modeling. Conceptual background
and Old World case studies. Parthenon, Carnforth, UK.
Chhatre, A., and A. Agrawal. 2008. Forest commons and
local enforcement. Proceedings of the National Academy
of Sciences USA 105:1328613291.
Churchill, S. P., H. Balslev, E. Forero, and J. L. Luteyn, editors.
1995. Biodiversity and conservation of neotropical montane
forests. New York Botanical Garden, Bronx, New York, USA.
Clewell, A., J. Aronson, and K. Winterhalder. 2004. The
SER International primer on ecological restoration. http://
www.ser.org/resources/resources-detail-view/ser-international-
primer-on-ecological-restoration
Colwell, R. K. 2009. EstimateS: statistical estimation of species
richness and shared species from samples. Version 8.2. Users
guide. Department of Ecology & Evolutionary Biology,
University of Connecticut, Storrs, Connecticut, USA.
Colwell, R. K., A. Chao, N. J. Gotelli, S. Y. Lin, C. X.
Mao, R. L. Chazdon, and J. T. Longino. 2012. Models
and estimators linking individual-based and sample-based
rarefaction, extrapolation, and comparison of assemblages.
Journal of Plant Ecology 5:321.
DAmico, L. 2010. Grassroots environmentalism in Intag,
Ecuador. Second UNITAR-Yale 2010 Conference on
Environmental Governance and Democracy, 1719
September 2010, New Haven, Connecticut, USA. http://
216 S. J. WILSON AND J. M. RHEMTULLA
conference.unitar.org/yale/sites/conference.unitar.org.yale/
files/Paper_Damico_MK_0.pdf
de la Pea-Domene, M., C. Martnez-Garza, and H. F. Howe.
2013. Early recruitment dynamics in tropical restoration.
Ecological Applications 23(5):11241134.
Dent, D. H., and J. S. Wright. 2009. The future of tropical
species in secondary forests: a quantitative review. Biological
Conservation 142:28332843.
Diemont, S. A., J. L. Bohn, D. D. Rayome, S. J. Kelsen,
and K. Cheng. 2011. Comparisons of Mayan forest
management, restoration, and conservation. Forest Ecology
and Management 261:16961705.
Ellis, E., and N. Ramankutty. 2008. Putting people in the
map: anthropogenic biomes of the world. Frontiers in
Ecology and Environment 2008:439447.
Fang, W., and S. L. Peng. 1997. Development of species
diversity in the restoration process of establishing a tropical
man-made forest ecosystem in China. Forest Ecology and
Management 99:185196.
Florentine, S. K., and M. E. Westbrooke. 2004. Restoration
on abandoned tropical pasturelands: do we know enough?
Journal for Nature Conservation 12:8594.
Freiberg, M., and E. Freiberg. 2000. Epiphyte diversity and
biomass in the canopy of lowland and montane forests in
Ecuador. Journal of Tropical Ecology 16:673688.
Gandolfi, S., S. V. Martins, and R. R. Rodrigues. 2007.
Forest restoration: many views and objectives. Pages 226
in R. R. Rodrigues, S. V. Martin, and S. Gandolfi,
editors. High diversity forest areas: methods and projects
in Brazil. Nova Science Publishers, New York, New York,
USA.
Garen, E. J., K. Saltonstall, J. L. Slusser, S. Mathias, M.
S. Ashton, and J. S. Hall. 2009. An evaluation of farmers
experiences planting native trees in rural Panama:
implications for reforestation with native species in
agricultural landscapes. Agroforestry Systems 76:219236.
Gentry, A. H. 1989. Northwest South America (Colombia,
Ecuador and Peru). Floristic inventories of tropical
countries. New York Botanical Garden, Bronx, New York,
USA.
Griscom, H. P., B. W. Griscom, and M. S. Ashton. 2009.
Forest regeneration from pasture in the dry tropics of
Panama: effects of cattle, exotic grasses, and forest riparia.
Restoration Ecology 17:117126.
Guariguata, M. R., and R. Ostertag. 2001. Neotropical
secondary forest succession: changes in structural and
functional characteristics. Forest Ecology and Management
148:185206.
Harris, J. A., and R. van Diggelen. 2006. Ecological restoration
as a project for global society. Pages 315 in J. van Andel,
and J. Aronson, editors. Restoration ecology. Blackwell,
Malden, Massachusetts, USA.
Harvey, C. A. 2000. Windbreaks enhance seed dispersal into
agricultural landscapes in Monteverde, Costa Rica.
Ecological Applications 10:155173.
Henderson, A., S. P. Churchill, and J. L. Luteyn. 1991.
Neotropical plant diversity. Nature 351:2122.
Higgs, E. 2003. Nature by design: people, natural process,
and ecological restoration. MIT Press, Cambridge,
Massachusetts, USA.
Hobbs, R. J., E. Higgs, and J. A. Harris. 2009. Novel
ecosystems: implications for conservation and restoration.
Trends in Ecology and Evolution 24:599605.
Hoch, L., B. Pokorny, and W. De Jong. 2012. Financial
attractiveness of smallholder tree plantations in the Amazon:
bridging external expectations and local realities.
Agroforestry Systems 84:361375.
Ecological Applications
Vol. 26, No. 1
Holl, K. D. 1998. The role of bird perching structures in
accelerating tropical forest recovery. Restoration Ecology
6:253261.
Holl, K. D. 1999. Factors limiting tropical rain forest regeneration
in abandoned pasture: seed rain, seed germination,
microclimate, and soil. Biotropica 31:229242.
Holl, K. D. 2002. Long-term vegetation recovery on reclaimed
coal surface mines in the eastern USA. Journal of Applied
Ecology 39:960970.
Holl, K. D., and T. M. Aide. 2011. When and where to
actively restore ecosystems? Forest Ecology and
Management 261:15581563.
Holl, K. D., M. E. Loik, E. H. V. Lin, and I. A. Samuels.
2000. Tropical montane forest restoration in Costa Rica:
overcoming barriers to dispersal and establishment.
Restoration Ecology 8:339349.
Holl, K. D., and R. A. Zahawi. 2014. Factors explaining
variability in woody above-ground biomass accumulation
in restored tropical forest. Forest Ecology and Management
319:3643.
Jacob, A. 2014. Conserving and restoring biodiversity and
ecosystem services in African tropical rainforest.
Dissertation. McGill University, Montreal, Quebec, Canada.
Jokisch, B. D., and B. M. Lair. 2002. One last stand? Forests
and change on Ecuadors eastern cordillera. Geographical
Review 92:235256.
Jost, L. 2013. Biogeography of the Pastaza watershed. http://
www.loujost.com/Biogeography%20of%20the%20
Pastaza%20Watershed/PastazaPrefaceIndex.htm
Kappelle, M., G. Avertin, M. E. Jurez, and N. Zamora.
2000. Useful plants within a campesino community in a
Costa Rican montane cloud forest. Mountain Research
and Development 20:162171.
Kirby, K. R., and C. Potvin. 2007. Variation in carbon
storage among tree species: implications for the management
of a small-scale carbon sink project. Forest Ecology and
Management 246:208221.
Klanderud, K., H. Z. H. Mbolatiana, M. N. Vololomboahangy,
M. A. Radimbison, E. Roger, . Totland, and C.
Rajeriarison. 2010. Recovery of plant species richness and
composition after slash-and-burn agriculture in a tropical
rainforest in Madagascar. Biodiversity and Conservation
19:187204.
Knowles, O. H., and J. A. Parrotta. 1995. Amazonian forest
restoration: an innovative system for native species selection
based on phenological data and performance indices.
Commonwealth Forestry Review 74:230243.
Kocian, M., D. Batker, and J. Harrison-Cox. 2011. An
ecological study of Ecuadors Intag region: the environmental
impacts and potential rewards of mining. Earth Economics,
Tacoma, Washington, USA.
Kozlowski, T. T. 1999. Soil compaction and growth of woody
plants. Scandinavian Journal of Forest Research
14(6):596619.
Kumar, N., A. Vajja, B. Pozzoni, and G. G. Woodall. 2005.
The effectiveness of World Bank support for community-
based and -driven development: an OED evaluation. World
Bank, Washington, D.C., USA.
Lamb, D., P. D. Erskine, and J. A. Parrotta. 2005. Restoration
of degraded tropical forest landscapes. Science 310:16281632.
Leopold, C. A., and J. Salazar. 2008. Understory species
richness during restoration of wet tropical forest in Costa
Rica. Ecological Restoration 26:2226.
Letcher, S. G., and R. L. Chazdon. 2009. Rapid recovery
of biomass, species richness, and species composition in a
forest chronosequence in northeastern Costa Rica.
Biotropica 41:608617.
January 2016 COMMUNITY-BASED CLOUD FOREST RESTORATION
Lovett, G. M., D. A. Burns, C. T. Driscoll, J. C. Jenkins,
M. J. Mitchell, L. Rustad, and R. Haeuber. 2007. Who
needs environmental monitoring? Frontiers in Ecology and
the Environment 5:253260.
Lugo, A. E. 1997. The apparent paradox of reestablishing
species richness on degraded lands with tree monocultures.
Forest Ecology and Management 99:919.
Lugo, A. E. 2009. The emerging era of novel tropical forests.
Biotropica 41:589591.
Lugo, A. E., and E. Helmer. 2004. Emerging forests on
abandoned land: Puerto Ricos new forests. Forest Ecology
and Management 190:145161.
Manzi, M., and O. T. Coomes. 2009. Managing Amazonian palms
for community use: a case of aguaje palm (Mauritia flexuosa)
in Peru. Forest Ecology and Management 257:510517.
McKinney, M. L., and J. L. Lockwood. 1999. Biotic
homogenization: a few winners replacing many losers in
the next mass extinction. Trends in Ecology & Evolution
14:450453.
Menninger, H. L., and M. A. Palmer. 2006. Restoring
ecological communities: from theory to practice. Pages
88112 in D. A. Falk, M. A. Palmer, and J. B. Zedler,
editors. Foundations of restoration ecology. Island Press,
Washington D.C., USA.
Michon, G., H. De Foresta, P. Levang, and F. Verdeaux.
2007. Domestic forests: a new paradigm for integrating
local communities forestry into tropical forest science.
Ecology and Society 12:1.
Mulligan, M. 2010. Modelling the tropics-wide extent and
distribution of cloud forest and cloud forest loss, with
implications for conservation priority. Pages 1438 in L.
A. Bruijnzeel, F. N. Scatena, and L. S. Hamilton, editors.
Tropical montane cloud forests: science for conservation
and management. Cambridge University Press, Cambridge,
Massachusetts, USA.
Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. da
Fonseca, and J. Kent. 2000. Biodiversity hotspots for
conservation priorities. Nature 403:853858.
Norden, N., R. L. Chazdon, A. Chao, Y. H. Jiang, and B.
Vlchez-Alvarado. 2009. Resilience of tropical rain forests:
tree community reassembly in secondary forests. Ecology
Letters 12:385394.
Ortega-Pieck, A., F. Lpez-Barreraa, N. Ramrez-Marcial,
and J. G. Garca-Francoa. 2011. Early seedling establishment
of two tropical montane cloud forest tree species: the role
of native and exotic grasses. Forest Ecology and
Management 261:13361343.
Parrotta, J. A. 1993. Secondary forest regeneration on degraded
tropical lands: the role of plantations as foster ecosystems.
Pages 6373 in H. Lieth, and M. Lohmann, editors.
Restoration of tropical forest ecosystems. Kluwer Academic
Publishers, Dordrecht, Netherlands.
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
6:627636.
Parrotta, J. A., J. W. Turnbull, and N. Jones. 1997. Catalyzing
native forest regeneration on degraded tropical lands. Forest
Ecology and Management 99:17.
Pedraza, R. A., and G. Williams-Linera. 2003. Evaluation of
native tree species for the rehabilitation of deforested areas
in a Mexican cloud forest. New Forests 26:8399.
Posada, J. M., T. M. Aide, and J. Cavelier. 2000. Cattle
and weedy shrubs as restoration tools of tropical montane
rainforest. Restoration Ecology 8:370379.
Rhemtulla, J. M., D. J. Mladenoff, and M. K. Clayton.
2007. Regional land-cover conversion in the U.S. Upper
217
Midwest: magnitude of change and limited recovery
in Wisconsin (185019351993). Landscape Ecology
22:5775.
Rhoads, C. C., G. E. Eckert, and D. C. Coleman. 1998.
Effect of pasture trees on soil nitrogen and organic matter:
implications for tropical montane forest restoration.
Restoration Ecology 6:262270.
Robbins, P. 2001. Tracking invasive land covers in India,
or why our landscapes have never been modern. Annals
of the Association of American Geographers
91:637659.
Ruiz-Jan, M. C., and T. M. Aide. 2005. Restoration success:
how is it being measured? Restoration Ecology
13:569577.
Sampaio, A. B., K. D. Holl, and A. Scariot. 2007. Regeneration
of seasonal deciduous forest tree species in long-used pastures
in Central Brazil. Biotropica 39:655659.
Sarmiento, F. O. 1995a. Restoration of equatorial Andes:
the challenge for conservation of Trop-Andean Landscapes
in Ecuador. Pages 637651 in S. P. Churchill, H. Balslev,
E. Forero, and J. L. Luteyn, editors. Biodiversity and
conservation of neotropical montane forests. New York
Botanical Garden, Bronx, New York, USA.
Sarmiento, F. O. 1995b. Human impacts on the cloud forests
of the Upper Guayllabamba river basin, Ecuador, and suggested
management responses. Pages 284295 in L. S. Hamilton, J.
O. Juvik, and F. N. Scatena, editors. Tropical montane cloud
forests. Springer-Verlag, New York, New York, USA.
Sarmiento, F. O. 1995c. Naming and knowing an Ecuadorian
landscape: the Maquipucuna Reserve as a case study. George
Wright Forum 12:1522.
Sarmiento, F. O. 2002. Anthropogenic change in the landscapes
of highland Ecuador. Geographical Review 92:213234.
Scatena, F. N., L. A. Bruijnzeel, P. Bubb, and S. Das.
2010. Setting the stage. Page 1 in L. A. Bruijnzeel, F. N.
Scatena, and L. S. Hamilton, editors. Tropical montane
cloud forests: science for conservation and management.
Cambridge University Press, Cambridge, Massachusetts,
USA.
Schneider, L. C. 2004. Bracken fern invasion in southern
Yucatan: A case for land-change science. Geographical
Review 94:229241.
Stattersfield, A. J., M. J. Crosby, A. J. Long, and D. C.
Wege. 1998. Endemic bird areas of the world. Birdlife
International, Cambridge, UK.
Terborgh, J., R. B. Foster, and P. Nuez. 1996. Tropical
tree communities: a test of the nonequilibrium hypothesis.
Ecology 77:561567.
Uhl, C., R. Buschbacher, R., and E. A. S. Serrao, . 1988.
Abandoned pastures in eastern Amazonia. I. Patterns of
plant succession. Journal of Ecology 76:663681.
Uprety, Y., H. Asselin, Y. Bergeron, F. Doyon, and J. F.
Boucher. 2012. Contribution of traditional knowledge to
ecological restoration: practices and applications. Ecoscience
19:225237.
Wassenaar, T., P. Gerber, P. H. Verburg, M. Rosales, M.
Ibrahim, and H. Steinfeld. 2007. Projecting land use changes
in the Neotropics: the geography of pasture expansion into
forest. Global Environmental Change 17:86104.
Wilson, S. 2013. Making tree projects work for the rural
poor in Latin America. Research to practice: strengthening
contributions to evidence-based policymaking. ISID-CIDA,
Montreal, Quebec, Canada.
Wilson, S. J. 2014. Replanting a future: restoring cloud
forests, biodiversity, and rural livelihoods in Andean
landscapes. Dissertation. McGill University, Montreal,
Quebec, Canada.
218 S. J. WILSON AND J. M. RHEMTULLA Ecological Applications
Vol. 26, No. 1

Wunder, S. 2007. The efficiency of payments for environmental and conservation of neotropical montane forests. New York
services in tropical conservation. Conservation Biology Botanical Garden, Bronx, New York, USA.
21:4858. Zahawi, R., and C. K. Augspurger. 1999. Early plant succession
Young, K. R., and B. Leon. 1995. Connectivity, social actors, in abandoned pastures in Ecuador. Biotropica
and conservation policies in the central Andes: the case 31:540552.
of Perus montane forests. Pages 653662 in S. P. Churchill, Zorrilla, C. 2010. Guia de reforestacion para las comunidades
H. Balslev, E. Forero, and J. L. Luteyn, editors. Biodiversity de Intag. http://www.decoin.org/currentwork/

SUPPORTING INFORMATION
Additional supporting information may be found in the online version of this article at http://onlinelibrary.wiley.com/
doi/10.1890/14-2129.1/suppinfo

DATA AVAILABILITY
Data associated with this paper have been deposited in Dryad: http://dx.doi.org/10.5061/dryad.41356