Biodivers Conserv (2014) 23:2183–2200

DOI 10.1007/s10531-014-0714-6

ORIGINAL PAPER

The abundance of rare tree species in remnant forests

across the Visayas, Philippines

Dennis Peque • Dirk Hölscher

Received: 27 November 2013 / Revised: 11 March 2014 / Accepted: 3 May 2014 /

Published online: 21 May 2014
Ó Springer Science+Business Media Dordrecht 2014

Abstract In several tropical regions of the world, formerly widespread forests have been
reduced to scattered remnants, and many tree species are becoming increasingly endan-
gered. Knowledge on the population status of rare species is essential to inform conser-
vation efforts and, in particular, the use of native species for reforestation projects
stemming from emerging land-use strategies. We studied 20 mostly red-listed native
species in remnant forests on five limestone-, four volcanic- and one ultramafic
site(s) across the Visayas in the Philippines, with 40 plots being assessed at each site.
Seventeen of the species showed a median density of less than two trees per hectare, with
ten species showing median densities of zero due to low species frequency, although some
had higher local densities. One species (Dracontomelon edule) was not encountered at all,
while two other species (Dipterocarpus validus and Dracontomelon dao) were only
encountered as singletons. Six species were confined to limestone sites and showed
associations with other better-known limestone specialists. The other study species
occurred at both site types. Four of the species showed significant relationships between
stem density, soil pH and stand basal area. The particular rarity of the majority of the
species calls for immediate conservation measures to be adopted to protect species and
associated remnant forests. In addition to legal measures such as the designation of rem-
nant forests as protected areas, we recommend the establishment of mixed native forest

Communicated by Peter Ashton.

Electronic supplementary material The online version of this article (doi:10.1007/s10531-014-0714-6)

contains supplementary material, which is available to authorized users.

D. Peque
College of Forestry and Environmental Science, Visayas State University, Visca, Baybay 6521, Leyte,
Philippines

D. Peque (&)
D. Hölscher

Tropical Silviculture and Forest Ecology, Georg-August-Universität Göttingen, Büsgenweg 1,
37077 Göttingen, Germany
e-mail: [email protected]

123
2184 Biodivers Conserv (2014) 23:2183–2200

stands giving due regard to the species associations and site characteristics identified in this
study.

Keywords Tropical forest
Limestone
Volcanic sites
Conservation

Red list
Reforestation

Introduction

In many tropical regions, old-growth forests only exist as scattered remnants, and many
tree species are becoming increasingly rare and endangered. Some formerly more abundant
tree species, such as those specific to lowland rainforest, may have become restricted to
remnant forest patches due to habitat loss. Other formerly rare species from the natural
continuous forest may have become critically endangered as a result of the deforestation
and fragmentation. Assessment of the current population status of rare tree species in
remnant forests has been scarce (e.g. Hubbell 2013; Bawa and Ashton 1991) and there is a
dearth of knowledge on associations among tree species and their relationships with site
conditions. Such information is helpful both to conservation and land management, and
particularly so as recent concepts of reforestation emphasise the use of native species and
tree species mixtures (Nguyen et al. 2012; Petit and Montagnini 2006).
Remnant forests may still support characteristic assemblages of species (Lucey et al.
2014), which may also include a large proportion of rare species or species with small
population densities (Hubbell 2013; ter Steege et al. 2013). As such, there is a definite
urgency to identify which species are rare, how they differ from one another and where
they can be found (Kaye et al. 1997). Species can be rare in many different ways. Rabi-
nowitz (1981) defined seven types of rarity based on: geographic range (wide or narrow),
habitat specificity (wide or narrow) and local population size (large or small). Species are
rare when they occur at low abundance or if they are restricted to specific habitats and have
narrow ecological ranges. Identifying the rarity features of species and knowing their
distribution in remnant forests is important for their conservation.
Vegetation studies in old-growth forests of the Philippines have thus far focused on
specific sites and are still limited in number (e.g. Hamann et al. 1999; Langenberger et al.
2006; Fernando et al. 2009), and in some cases, only the presence of dominant or com-
mercially important species is emphasized (Cao et al. 2001). While such studies are
important, there remains a need for more knowledge to be developed on rare and lesser
studied species at a larger geographical scale and at different sites. In the Visayas, par-
ticularly in Leyte, the so-called ‘‘rainforestation approach’’ to reforestation has been
adopted, which involves planting many native species in mixture, and some interesting
stands have already been developed (Milan and Margraf 1994; Nguyen et al., 2012,
Schneider et al. 2014). The success of the scheme offers encouragement to wider areas and
plays a prominent role in the nationwide concept of conservation. Access to planting
material has however been one of the major constraints and concerns have been raised over
the provenance or genetic diversity of seedlings used, as they may come only from a single
or very limited number of mother trees (pers. Com. Fernandez). Thus, surveys on tree
species that represent potential sources of seed, especially in forest remnants, are deemed
important in this regard.

123
Biodivers Conserv (2014) 23:2183–2200 2185

Many recent studies have investigated the importance of soil nutrients, topography and
climate in shaping tree species communities (Bohlman et al. 2008; Condit et al. 2013;
Clark et al. 1998). However, the high diversity of tropical forests constituted by an
extraordinarily large share of rare species (Hubbell 2013; ter Steege et al. 2013) compli-
cates our understanding on the effects of environmental factors on species’ abundance and
distribution (Condit et al. 2013). For our study, we selected 20 tree species native to the
Visayas that are currently red-listed, many of which are being used in reforestation. Five of
these species are endemic to the Philippines and two (Vitex parviflora and Afzelia rhom-
boidea) are known to have specialised or preferred habitats (Whitford 1911; de Guzman
et al. 1986). Most of the selected species are formerly widespread and abundant, but they
are currently all assumed to be endangered due to habitat loss. Forty study plots were
selected from each of ten sites within remnant forests overlying soils derived from lime-
stone and non-limestone substrates (4 intermediate volcanic and 1 ultramafic). The
objectives of the study were to determine: (1) the abundance and distribution of the
selected species in the study sites; (2) which, if any, of the species showed associations
with other species and sites; and (3) the environmental factors related to the presence and
abundance of each study species. The study shall provide basic information on the current
population status of selected tree species and the factors that govern their distribution,
which will be essential to the management and conservation of both species and remnant
forests.

Materials and methods

The Visayas region

The Visayas region encompasses the central group of islands in the Philippine archipelago
and comprises an estimated land area of 57,000 km2. Rugged terrain and high mountains
characterize most of the Visayan Islands and limestone and volcanic deposits can be found
throughout. The climate in the Visayan lowlands is tropical and monsoonal with an average
air temperature of 27.9 °C and little seasonal variation. Annual rainfall reaches
?4,000 mm (4,600 mm year-1 recorded in Negros, Hamann et al. 1999) at higher ele-
vations and falls to 1,000 mm at leeward (rain shadow) sites and in sheltered valleys. The
wet season usually runs from May to October and, where occurring, the dry season from
November to April.
Original vegetation on coastal hills and uplands in the Visayas comprises evergreen
tropical rainforest. Forest on volcanic soils features a closed canopy dominated by di-
pterocarps, whereas limestone forests are relatively open and interspersed with only a few
dipterocarps (Whitford 1911; Fernando et al. 2008). Like other parts of the Philippines, the
Visayas have experienced extensive deforestation. Records of harvesting premium tree
species in the Visayas date back to the Spanish colonial period (Bankoff 2007). Modern
logging operations started in the early 1900s on the island of Negros, but many of the
forests in the Visayas were lost during the 1960s, until the implementation of logging
regulation in 1991. Currently, most of the primary lowland and hill dipterocarp forests
have been replaced with coconut plantations, agricultural farms or degraded open lands
covered with grasses and a few pioneer tree species.

123
2186 Biodivers Conserv (2014) 23:2183–2200

Fig. 1 Location of study sites in the Visayas, Philippines (solid triangles = forests over limestone (FOL);
open triangles = forests on non-limestone (FNL) substrates)

Study sites

The study was carried out across ten different forest areas on six islands in the Visayas
(Fig. 1; Supplementary Table 1). The sites were selected following consultation with local
experts and forestry offices, which have jurisdiction over certain forest areas. The criteria
used in the selection process were: (1) presence of forests with a minimum size of 30
hectares; (2) accessibility, and (3) in accordance with health and safety protocols. Remnant
forests in the Visayas are limited and our study sites did not represent truly undisturbed
primary forest, as signs of human encroachment were evident. The management and
protection of these forests rests on the Department of Environment and Natural Resources
(DENR), and specifically, the Community Environment and Natural Resources Office
(CENRO). In some areas, local government units (LGUs) and people’s organisations (POs)
support DENR in the management and protection of these remnant forests. Forest areas in
the region where tree inventories had previously been conducted, such as Mount Pan-
gasugan (Langenberger et al. 2006) and Negros (Hamann et al. 1999), were not re-visited
but served as comparison.
The study sites covered lowlands up to the transition zone between lowlands and lower
montane forests, which are believed to be potential refugia for endangered tree species.
Floristic composition in the transition zone forest (Sibulan site) bears some resemblance to
other forests located at lower montane elevations (e.g. [800 m asl; Almeria, Biliran), and
members of family Fagaceae are common to the sites. The selected study sites covered five
forests on limestone soils, four on volcanic soils and one in ultramafic soil (Salcedo site).
Here, we designate our study sites as: ‘forest over limestone’ (FOL) and ‘forest on non-
limestone’ (FNL) substrate. Volcanic and limestone substrates are the most widespread and
dominant in the Philippines. The classification and nomenclature of forests in the Philip-
pines as proposed by Fernando et al. (2008) is based on the main physical characteristics of
their habitat, vegetation structure and physiognomy and not solely on the kind of soils

123
Biodivers Conserv (2014) 23:2183–2200 2187

where the forests are growing. In the case of Whitford (1911), he broadly classified the
forests in the Philippines based on stand composition. Both systems include forest types
occurring over limestone (Whitford refers to ‘Molave forest’ as forest on limestone
dominated by V. parviflora,), but neither includes a classification for ‘forests on volcanic
soils’, even though many forest types can occur on this soil. Such forests may have
survived particularly well until relatively recently due to them being formerly inaccessible
and away from local communities, while a few others (e.g. Bilar Bohol) have benefited
from being designated within protected areas for some time.
The limestone sites were generally characterised by a karst topography, mainly com-
prising sharp rocky outcrops, some sinkholes and very rugged terrain. The soils were
generally shallow at around 10–15 cm deep with colours ranging from black to greyish
black. Dark brown, brown or reddish-brown soils were occasionally observed in pockets.
The elevation of the limestone study sites ranged from 207 to 776 m asl with gentle to
moderate slopes. The non-limestone sites were represented by four intermediate volcanic
soils (Jahn and Asio 1998) and by a special type of soil (Ferralsols), which is ultramafic in
composition (Navarrete et al. 2007; Travaglia et al. 1978) and derived from ophiolites
(Dimalanta et al. 2006). All the non-limestone sites had deeper soils that extended to one
meter or more in depth, with the exception of some sampling locations in the Sibulan and
Ormoc sites that had thin soil layers on top of loose, partially weathered stones or rocks.
Soil colour ranged from dark brown to yellowish or reddish brown to lighter brown and
elevation at the sites range from 107 to 1,040 m asl, which did not compare well with that
of the limestone sites due to locational constraints.
Due to the lack of specific meteorological stations, some study sites share common
climate data. Annual rainfall nearest to the limestone sites ranged from 1,661 to 3,127 mm
per year, and from 1,408 to 3,437 mm per year for non-limestone sites (see supplementary
Table 1). The number of dry months in the study sites ranged from 2 to 4 months with
some areas experiencing no dry months, and the study sites located in the same climatic
zones had the same number of dry or wet months. There was little variation in annual mean
temperature.

The focal study tree species

We selected 20 tree species (Table 1) based on presumed rarity, potential use in refores-
tation and ease of identification. In the International Union for the Conservation of Nature
Red List (IUCN 2013), five of the species are categorized as critically endangered, one as
endangered and seven as vulnerable; the other seven are currently not evaluated. Thirteen
of the species are also listed in the Philippine Red List (DAO 2007-01). All but one are
promoted in Rainforestation farming (Milan and Margraf 1994) and partly belong to the
most preferred native tree species for smallholder forestry on Leyte, Philippines (Man-
gaoang and Pasa 2003). Of the 20 selected species, five species belong to the Diptero-
carpaceae, four to the Fabaceae (Leguminosae), two to the Verbenaceae, two to the
Anacardiaceae and one each to the families Moraceae, Tiliaceae, Clusiaceae, Meliaceae,
Ebenaceae, Myrtaceae and Sapotaceae.

Assessment of focal species and forest tree inventory

For each study site, we established 40 plots of 400 m2 in size following a systematic
sampling approach with a random starting point. The sampling plots were arranged in grids
of 5 9 8, giving a total of 40 plots. The plots were spaced at 100 m from each other and

123
 Table 1 The focal tree species and their conservation status
2188

Scientific name Family name Official Species Conservation status Use in Distributionb
common code refores-
Philippine IUCN

123
name tation
red lista 2013

Shorea almon Foxw. Dipterocarpaceae Almon Salm VU CR Yes Borneo; widespread in the Philippines
Parashorea malaanonan Dipterocarpaceae Bagtikan Pmal NE CR Yes Borneo, East Kalimantan, Philippines; lowlands up to 1,300 m
(Blanco) Merr. asl
Dipterocarpus validus Dipterocarpaceae Hagakhak Dval NE CR Yes Sabah, Kalimantan, Philippines; flatlands and riverine sites and
Blume hills up to 300 m asl
Shorea palosapis Dipterocarpaceae Mayapis Shopa NE CR Yes Philippine endemic; on fertile well-drained soil, often abundant
(Blanco) Merr.
Shorea contorta Vidal Dipterocarpaceae White Lauan Shoco VU CR Yes Philippine endemic; lowlands up to 700 m asl; widespread
Wallaceodendron Fabaceae Banuyo Wacel EN NE Yes Sulawesi and Philippines; lowlands up to 850 m asl
celebicum Koord.
Intsia bijuga (Colebr.) Fabaceae Ipil Ibij VU VU Yes Indo-Pacific; along seashore up to 600 m asl
Kuntze.
Pterocarpus indicus Fabaceae Narra Pind CR VU Yes Southeast and East Asia, northern and southwest Pacific region;
Willd. at low and medium altitudes
Afzelia rhomboidea Fabaceae Tindalo Arho EN VU Yes Sumatra, Borneo and Philippines; at low and medium altitudes
(Blanco) Vidal
Artocarpus blancoi Moracaeae Antipolo Abla NE VU Yes Philippine endemic; widely distributed
(Elmer) Merr.
Diplodiscus paniculatus Tiliaceae Balobo Dpan NE VU Yes Philippine endemic; occur at low and medium altitudes
Turcz.
Calophyllum blancoi Clusiaceae Bitanghol Cabla NE NE Yes Lanyu Island, Philippines and Borneo; in lowland forests up to
Planch. &Triana 1,900 m asl
Vitex quinata (Lour.) F. Verbenaceae Kulipapa Vqui NE NE No Indonesia, India, China, Malaysia, Thailand, Philippines;
N. Will.
Vitex parviflora Juss. Verbenaceae Molave Vipar EN VU Yes Indonesia, Philippines; tolerates a wide range of soils but
occurs mostly on dry limestone soils
Biodivers Conserv (2014) 23:2183–2200
 Table 1 continued

Scientific name Family name Official Species Conservation status Use in Distributionb
common code refores-
name Philippine IUCN tation
red lista 2013

Dracontomelon dao Anacardiaceae Dao Ddao VU NE Yes India, Myanmar, Thailand, Cambodia, S. China, scattered
(Blanco) Merr.& Rolfe throughout Malesia and Solomon Isls.; wide habitat; occur
from near sea-level to 500–1,000 m asl
Dracontomelon edule Anacardiaceae Lamio Dedu VU NE Yes No information
Merr.
Toona calantas Merr. & Meliaceae Kalantas Tcal CR NE Yes Sumatra, Peninsular Malaysia, Borneo, Java, Guinea;
Rolfe widespread in the Philippines
Biodivers Conserv (2014) 23:2183–2200

Diospyros philippinensis Ebenaceae Kamagong Dphi CR EN Yes Sulawesi, Taiwan; widespread in the Philippines
A. DC.
Tristaniopsis decorticata Myrtaceae Malabayabas Tdec CR NE Yes Philippines and Sulawesi; occur at low and medium altitude
(Merr.) Peter G. Wilson
& J.T. Waterh.
Palaquium luzoniense Sapotaceae Nato Pluz VU VU Yes Philippine endemic; widely distributed
(Fern.-Vill.) Vidal
CR means critically endangered; EN endangered; VU vulnerable; and NE not evaluated
a
Based from DAO 2007-01
b
Mainly based from Newman et al. (1996), de Guzman et al. (1986), Lemmens et al. (1995)
2189

123
2190 Biodivers Conserv (2014) 23:2183–2200

were laid following the altitudinal contour. Slope correction was applied for each plot and
distance between plots. On each plot, all trees belonging to the focal species with a
diameter at breast height (dbh) of C10 cm were recorded and their corresponding diam-
eters measured.
Of the 40 plots, five were randomly selected to collect further stand structure and soil
data, with all trees of dbh C10 cm being identified as far as possible and recorded.
Unidentified tree species on these plots were coded (i.e. species A, B, C etc.) and leaf
samples were collected to compare unidentified species from other plots within each study
site. Tree heights were measured using a vertex hypsometer (Vertex IV, Haglöf, Sweden).
Tree height was estimated visually in cases where topographic features rendered it too
difficult to measure.

Soil sampling and analysis

Samples from the mineral soil were collected at 0–10 cm, and colour and texture were
examined during sampling. Soil samples were then air-dried and pulverized and passed
through a 2-mm sieve. Analyses were performed at the Visayas State University Soils
Research Testing and Plant Analysis Laboratory (SRTPAL) Baybay, Leyte, Philippines,
and they followed the standard procedure of ISRIC (1995): soil pH in water was measured
using glass electrodes in a supernatant with 1:2.5 soil/water ratio, organic matter content
(OM; Walkley–Black method), total nitrogen (N; Kjeldahl), extractable Phosphorus (P;
Olsen), and cation exchange capacity (CEC; ammonium acetate method).

Data analysis

In order to find possible associations between species, reciprocal averaging (RA) of focal
species and dominant tree species was carried out using PC-ORD software (version 5.12;
McCune and Mefford 2006). For each study site, one species was selected as being
dominant based on the respective importance value index attained (Curtis and McInstosh
1951). By combining density data of the 19 focal species (one species was not encountered
in any of the study sites) with that of the most dominant tree species on each site, a
10 9 24 species matrix was formed. The data was normalized by logarithmic transfor-
mation prior to running RA.
To investigate the relationships between densities of focal tree species and environ-
mental and stand structural variables, a detrended correspondence analysis (DCA) was
conducted, also using PC-ORD. Matrix 1 consisted of tree density data for each of the 19
target species identified. Matrix 2 utilized the mean values of ten quantitative variables
including soil pH, N, P, SOC, CEC, elevation, tree stand heights, top heights (i.e. the height
of the tallest tree in each inventory plot), stand basal area, stem density per ha of all trees
and one categorical variable (FOL vs. FNL). Because data were not normally distributed,
Matrix 1 was subjected to logarithmic transformation, and data in Matrix 2 were stan-
dardized to ensure that an equal weighting was attributed to each variable. Twenty-six
segments and a rescaling threshold of zero were used, and the cut-off r2 value was set to
0.45. Pearson correlations were calculated to check patterns indicated by DCA. Initial
results showed that soil pH, cation exchange capacity (CEC), stand basal area (BA) and
total stem density were the only independent variables among the ten variables that were
significantly correlated with the DCA axes. Since soil pH and CEC were significantly
correlated, the former was omitted in the next analysis. Stem density was also excluded
because it also correlated with basal area.

123
Biodivers Conserv (2014) 23:2183–2200 2191

Similarity between study sites was determined using a Morisita-Horn index (Wolda
1981) based on the density of focal species. The resulting values were used in running a
Mantel test to determine the relationship between geographical distance and density of
focal species. Euclidian distance was used for Matrix 1 and Sörensen (Bray-Curtis) for
Matrix 2 using a Monte Carlo test with 999 randomized runs.

Results

General stand and site characteristics

Eight tree species constituted the most dominant species across the study sites, with two
species (Shorea palosapis and Shorea astylosa) being dominant in two of the sites. Among
the dominant species, four were dipterocarps (Shorea contorta, Shorea polysperma, S.
palosapis, and S. astylosa) while the others belong to the family Podocarpaceae, Fabaceae
(Leguminosae), Casuarinaceae and Sapotaceae.
Stands on limestone (FOL) sites had statistically significant lower basal areas than the
forest stands on non-limestone (FNL) sites, and the top heights tended to be smaller
(Table 2). Diameter, stem density, estimated total numbers of tree species and Shannon H’
and Evenness were more similar. Very few trees reached C100 cm DBH.
There were significant differences in soil chemical properties between FOL and FNL
sites except for extractable phosphorus (P) (Table 2). FOL sites had higher mean soil pH,
soil organic carbon (SOC), total nitrogen (N) and cation exchange capacity (CEC) values
compared to FNL sites.

Occurrence and density of focal tree species

One species, Dracontomelon edule, was not encountered at all, while four other species
(Intsia bijuga, Diplodiscus paniculatus, Afzelia rhomboidea and Dipterocarpus validus)
were each found to occur only on a single site (Supplementary Table 2). The species D.
validus and Dracontomelon dao appeared as singletons on one and two study sites,
respectively. Due to the low species frequency, ten of the focal species showed a median
density of 0 per ha, one species was 1.0 per ha, another was 1.5 per ha, and three species
had a median density of 0.5 trees per ha (Fig. 2). Local stem densities were very low (\5
trees per ha) for the majority of species, and only a few reached [10 trees per ha. On FOL
sites, only five species had a recorded density of C10 trees per ha on at least one site, while
there were only six species that attained the same density on FNL sites. The most wide-
spread species were Calophyllum blancoi and Palaquium luzoniense, which occurred
across nine and all sites, respectively. There were significant differences in density in 12 of
the focal species between FOL and FNL study sites (Table 3).

Associations among species and sites

All focal species encountered (19) were found on limestone sites while only 13 were
recorded in forests on non-limestone substrates (Supplementary Table 2). The six focal
species that occurred exclusively on limestone sites were V. parviflora, A. rhomboidea, I.
bijuga, Wallaceodendron celebicum, D. validus and D. paniculatus. Four of these species
tended to associate with each other, and together with Pterocarpus indicus (which occurred

123
2192 Biodivers Conserv (2014) 23:2183–2200

Table 2 Soils and stand (all species) characteristics of the study sites in forests over limestone and forests
on non-limestone substrates
Variables Limestone Non-limestone

Median Min Max Median Min Max

Environment
pH (H2O) 5.7a 4.6 6.5 4.1b 3.8 5.0
-1 a
SOC (g kg ) 54.7 51.8 79.9 38.2b 33.6 71.8
N (g kg-1) 6.0a 5.2 7.9 3.5b 2.4 7.5
C/N ratio (g g-1) 9.2a 8.7 10.2 11.0b 8.4 14.2
Extractable P (mg kg-1) 4.6a 1.1 16.9 5.4a 0.7 37.3
CEC (me kg-1) 632a 348 826 242b 202 554
Elevation (m asl) 452a 207 776 779a 107 1040
Forest structure
DBH (cm) (C10 cm dbh) 23.0a 15.4 23.4 24.2a 22.1 27.4
a
Stem density (C10 cm dbh) (n per ha) 375 210 390 500a 345 725
Basal area per ha (m2) 17.5a 4.5 20.7 32.7b 22.7 49.1
Total tree heights (m) 15.7a 12.4 21.0 15.5a 10.9 19.9
Top heights (m) 18.2a 14.5 25.8 25.9a 19.5 27.7
Tree species richness (n per 2,000 m2) 40a 23 42 43a 39 44
Shannon H’ 3.43a 2.98 3.51 3.38a 3.3 3.65
Evenness (Shannon) 0.93a 0.93 0.95 0.90a 0.88 0.97
Differing letters indicate statistical difference (p B 0.05) in the Mann–Whitney test (n = 5)

Fig. 2 Frequency and median density of focal tree species in the ten study forests across the Visayas,
Philippines. Refer to Table 1 for species codes

on both soil types), they formed a distinct group, as shown in their arrangement in the RA
ordination plot (Fig. 3). Reciprocal averaging also resulted in the association of the four
other dominant species identified in FOL (S. astylosa—code SHOAS, S. contorta—
SHOCO, Casuarina rumphiana—CARUM and S. Palosapis—SHOPA) with all other tree
species in the FNL group except V. parviflora (VIPAR). A gradient of species along RA
axis 2 was recognizable, with Palaquium philippense (PALPH) having the highest score.

123
Biodivers Conserv (2014) 23:2183–2200 2193

Table 3 Density (n per ha) of focal tree species in forests over limestone and forests on non-limestone
substrates (n = 5)
Species Limestone Non-limestone

Median Min Max Median Min Max

Shorea almon 2a 0 8 1b 0 41
a
Parashorea malaanonan 0 0 3 1b 0 13
Dipterocarpus validus 0a 0 1 0a 0 0
Shorea palosapis 2a 0 12 19b 0 75
Shorea contorta 1a 0 27 1b 0 7
Wallaceodendron celebicum 0a 0 9 0b 0 0
Intsia bijuga 0a 0 19 0b 0 0
Pterocarpus indicus 0a 0 4 1a 0 1
Afzelia rhomboidea 0a 0 5 0b 0 0
Artocarpus blancoi 1a 0 7 3a 0 4
Diplodiscus paniculatus 0a 0 4 0b 0 0
Calophyllum blancoi 3a 0 23 8b 4 48
Vitex quinata 0a 0 1 4b 0 14
Vitex parviflora 0a 0 9 0b 0 0
Dracontomelon dao 0a 0 1 0a 0 1
Dracontomelon edule 0 0 0 0 0 0
Toona calantas 0a 0 3 2a 0 3
a
Diospyros philippinensis 1 0 3 0a 0 8
Tristaniopsis decorticata 1a 0 4 1a 0 5
Palaquium luzoniense 6a 1 19 11b 4 28
Differing letters indicate significant differences (p \ 0.05) in the Mann–Whitney test

No species were found to occur exclusively in forests on non-limestone sites. The studied
tree species in the ultramafic site in Salcedo were not particularly outstanding or different
from the other sites. Based on the Mantel test, no significant correlation between geo-
graphical distance and density of any given focal tree species was indicated.

Environmental factors influencing density of focal tree species

Eight variables (pH, N, P, SOC, elevation, BA, tree heights and top heights) were included
in the final DCA. The DCA axes explained a total inertia (variance) of 1.89 and produced
eigenvalues of 0.71 and 0.29 for the first and second axes respectively, which indicates that
axis 1 better explains the spread of species and independent variables than axis 2.
DCA axis 1 (Fig. 4) was positively correlated with soil pH (0.75, p \ 0.05) and neg-
atively correlated with stand basal area (0.75, p \ 0.05). The densities of I. bijuga (0.92),
A. rhomboidea (0.92), P. indicus (0.78) and V. parviflora (0.68) were correlated with axis 1
at p \ 0.01. This indicates that the densities of these species tend to increase with
increasing soil pH and tend to decrease with increasing stand basal area. However, four
species were also significantly correlated with axis 2 at p \ 0.05, and one among them (S.
contorta) correlated negatively. The density of some other species showed very weak
correlations with axis 1, indicating that they are less influenced by soil pH and basal area.

123
2194 Biodivers Conserv (2014) 23:2183–2200

Fig. 3 Position of focal species and dominant tree species as defined by the first two axes of reciprocal
averaging (RA). Eigenvalues: 0.57 (axis 1), 0.38 (axis 2); total variance: 2.7. The forest over limestone
(FOL) group comprises four species that occurred exclusively on limestone and one species, Pind
(Pterocarpus indicus), that occurred on both site types. The other group (FNL ? FOL) comprises the
remaining focal species and seven dominant species (upper case letters) across the study sites. Refer to
Table 1 for species codes

Fig. 4 DCA biplot of focal species as well as environmental and stand structure variables. Eigenvalues:
0.71 (axis 1), 0.29 (axis 2); total variance: 1.89. Solid triangles are forests over limestone (FOL), open
triangles are forests on non-limestone (FNL) substrates, while crosses represent tree species. Refer to
Table 1 for species codes

123
Biodivers Conserv (2014) 23:2183–2200 2195

Discussion

Ecological status of the focal species

Many of our tree species were found to be rare, but this may be expected for a typical
diverse tropical rainforest (Hubbell 2013; ter Steege ter Steege et al. 2013). The presence
of singletons, species with very low numbers of individuals, and species that appear only in
a single or in very few sites indicates that the majority of our focal species are currently
rare. For instance, D. edule was not encountered in the study sites but was observed by
other researchers (e.g. Langenberger et al. 2006; Cadiz and Buot 2010) at low abundance.
The species D. Dao was only encountered in two sites and as a singleton but was similarly
recorded only at \1 tree per ha in the Pasoh forest reserve in Malaysia (Saw et al. 1990).
These observations indicate that there are species in the Visayas whose populations are
already critically low that should be given immediate priority for conservation.
Some species that appeared locally rare were common or abundant elsewhere. The
higher density of I. bijuga (19 trees per ha) in Argao and its complete absence at other
study sites also suggests that this species should become a priority for conservation. The
species was only recorded at \1 tree per ha in a sub-montane forest in Negros, Visayas
(Turner et al. 2003) but at a very high density of 52 trees for a 0.2 ha transect in Mt.
Malarayat, Luzon (Lunar and Arcega 2011). The distribution and abundances of five
selected dipterocarps were also disproportionate, being high, low or absent in some study
sites, with the most common being S. palosapis, as also confirmed in other studies (e.g.
Schade et al. 1987).
The differences in abundance and occurrence of our focal species can be attributed to
many factors. Firstly, our sampling method was not designed to detect rare events (e.g.
adaptive cluster sampling (ACS), Thompson 1992; Acharya et al. 2000), such that it may
have failed to include some species or individuals in the sample, despite their possible
presence in the study sites. Secondly, the diameter limit of 10 cm dbh may be slightly
greater than ideal for a sampled area of 1.6 ha, in that it may have excluded smaller trees
(Kenfack et al. 2007). Forest fragment size and the level of disturbance (Ranta et al. 1998)
may also have contributed to the current population of the species in the study sites, and
anthropogenic activities may have reduced the density of some commercially important
species, especially in forests located outside protected areas. This could be the reason why
some study sites lacked larger canopy trees and, in particular, why some sites had low stem
densities. The significantly lower basal area of forests over limestone compared to the non-
limestone forests could be partly due to disturbance and partly to site conditions. The
limestone sites have a karst topography and shallow soils, which may have a profound
effect on water availability and that needed for tree growth (Estrada-Medina et al. 2013).
Nevertheless, our study areas ranged in the size from 30 ha up to more than 500 ha, and the
smaller fragments were already isolated, especially on most of the limestone sites. How-
ever, some of these smaller forest fragments contain some species not found in other larger
fragments, suggesting that the former may be of higher conservation value, similar to that
of the Otoville lowland forest of American Samoa which serves as an important refugium
for unique flora (Seamon et al. 2006).

Species associations

All 19 encountered focal species were present on limestone sites, and six of these were
exclusively confined to limestone. At undisturbed sites, limestone specialist species such as

123
2196 Biodivers Conserv (2014) 23:2183–2200

V. parviflora and A. rhomboidea would be expected to be found on limestone, although

they could also be encountered in other soil types (de Guzman et al. 1986; Langenberger
et al. 2006; Fernando et al. 2008). However, other species that are known to occur on a
wide range of soils or habitats, such as D. validus and D. paniculatus, were also
encountered only on limestone. It may be that the inclusion of these two species in the
sample on limestone was only by chance as their populations are currently critically small,
with D. validus being encountered only on a single site and as a singleton. Based on the
dominant species approach described by Whitford (1911) for the Molave forest type
(limestone), only four of the six species can be more assuredly considered limestone
specialists. This conclusion is further supported when considered in terms of the deciduous
character of the four species. Dry-season moisture or water stress on limestone soils is an
important predictor of species distributions (Condit et al. 2013). However, the association
of P. indicus with other better-known limestone specialists suggests that it is more of a
limestone species than a generalist. This finding is largely consistent with earlier studies of
Whitford (1911), which indicate that P. indicus may be dominant in some expressions of
the Molave forest type. All of our other focal species can be considered generalists, and the
dipterocarps may be expected to become dominant on non-limestone soils but may only
share dominance in forests over limestone substrate (Fernando et al. 2008). Legumes (P.
indicus, W. celebicum, A. rhomboidea, I. bijuga) constitute our limestone species, as has
similarly been observed for neotropical limestone forest elsewhere (Pennington et al.
2000).

The influence of soil nutrients and basal area on tree species density

Our study provides a little support to the notion that soil nutrients affect tree species
distribution, species composition, and species diversity (Valencia et al. 2004; Potts et al.
2002; Condit et al. 2013). Among the nutrients tested, we found soil pH to be positively
correlated with 21 % of the focal species, suggesting they are more sensitive to soil pH
(Goldberg 1985). The strong response of these focal species to limestone may highlight the
role of soil type on species’ abundance and distribution (Valencia et al. 2004), as supported
by the significant difference in soil pH between FOL and FNL sites. However, Clark et al.
(1998) argued that plants do not respond to soil type per se but to specific soil nutrients or
properties such as pH. Conversely, Sahu et al. (2012) found a negative correlation between
soil pH with both density and tree diversity in the tropical forest of Eastern Ghats, India. In
our study, the low correlation between species density and nutrients on FNL soils is
consistent with the results of Potts et al. (2002) on Bornean tropical rain forest. The
negative correlation of limestone specialists with basal area reflects the characteristic of
limestone forests in that they are more open and the trees are short-boled (Whitford 1911).
Smaller basal area could also be a result of the quality of site conditions (shallow soils,
water stress, etc.), especially in drier regions. In contrast, our data suggest that the trees
found in FNL had greater diameters and hence larger basal areas.

Unexplained environmental factors influencing density of species

Our results did not show any relationship between geographical distance and densities of
focal tree species, which departs from findings from studies in other regions of the world
on species over different geographical ranges (e.g. Potts et al. 2002 on Borneo; Bohlman
et al. 2008; Guéze et al. 2013, on Amazon). In addition, elevation did not show any
correlation with the ordination axes, since most of the species have wide elevation ranges

123
Biodivers Conserv (2014) 23:2183–2200 2197

such that they occur from lowlands up to the transition zone between lowlands and lower
montane forests. With regard to precipitation, it was excluded from our analysis due to the
lack of meteorological stations in the proximity of our study sites as well as the potential
low reliability of using WorldClim data (Hijmans et al. 2005) in view of the high
mountains and consequent effects of windward and leeward exposure.

Management considerations

This study provided information on factors that influence species distribution and the
current population status of some species, which may be useful in updating the Philippine
red list, as a baseline for designing appropriate sampling methods for the species or in
developing conservation measures both for the species and remnant forests. In the interests
of conducting future forest inventories, we suggest that appropriate sampling methods be
considered based on the current population densities for each species. Currently, the
majority of the focal species have modest local populations or are particularly rare and they
should be given immediate high conservation priority. However, we also found some
species that are still widespread and abundant in our study area, even so, we recommend
they be categorised as rare, as they are currently restricted to the remaining limited forest
patches in the Visayas.
The apparent absence of one species and the occurrence of two other species as sin-
gletons highlights a lack of conservation of the species across the sites. Nevertheless,
potential sources of planting materials of the majority of the species can be identified
within the study areas. We therefore advocate the establishment of a plantation of mixed
native species to increase existing populations of the focal species, and probably adopting
the Rainforestation approach (Milan and Margraf 1994; Schneider et al. 2014). Species
characteristics, observed associations among species and their suitability to site conditions
should be considered in the development of a planting design. Although reliable studies on
the effect of the collection of seeds and seedlings and its impact on forest dynamics from
the region are not available, we expect that limited harvesting of seeds and seedlings for
reforestation within and outside the remnant forests may be beneficial to the conservation
of rarer species. To this end, the origin of planting material should be documented,
especially in cases where obvious provenances exist, e.g. for Parashorea malaanonan
(Villarin 2013).
We also suggest that all study sites that are currently not under the status of National
Park or part of any declared critical watershed (i.e. Inopacan, Silago, Alcoy and Salcedo)
be converted into protected areas to provide legal protection for these forest remnants. In
designing management plans, the inclusion of socio-economic activities at buffer areas
should also be considered in order to minimise the pressure exerted by local populations
close to these remnant forests and to make them more active in helping manage and protect
the forests (Lynagh & Urich 2002). Although people’s organizations (POs) and some local
government units (LGUs) are involved in managing and protecting some of these forests,
their participation needs to be improved and strengthened.

Acknowledgments We thank Mr. Junar Lendio for his invaluable assistance in the conduct of this
research and the Deutsche Akademischer Austauschdienst (DAAD) for the scholarship received by the first
author.

123
2198 Biodivers Conserv (2014) 23:2183–2200

References

Acharya B, Bhattarai G, de Gier A, Stein A (2000) Systematic adaptive cluster sampling for the assessment
of rare tree species in Nepal. For Ecol Manag 137:65–73. doi:10.1016/S0378-1127(99)00318-7
Bankoff G (2007) One island too many: reappraising the extent of deforestation in the Philippines prior to
1946. J Hist Geogr 33:314–334. doi:10.1016/j.jhg.2006.06.021
Bawa KS, Ashton PS (1991) Conservation of rare trees in tropical rainforests. A genetic perspective. In: Falk
DA, Holsinger KE (eds) Genetics and conservation of rare plants. Oxford University Press, New York
Bohlman SA, Laurance WF, Laurance SG et al (2008) Importance of soils, topography and geographic
distance in structuring central Amazonian tree communities. J Veg Sci 19:863–874. doi:10.3170/2008-
8-18463
Cadiz GO, Buot IE Jr (2010) An enumeration of the vascular plants of Mt. Tabunan, Cebu Island, Phil-
ippines. Thail Nat Hist Mus J 4(2):71–77
Cao Y, Larsen DP, Thorne RSt-J (2001) Rare species in multivariate analysis for bio-assessment: some
considerations. J N Am Benthol Soc 20(1):144–153
Clark DB, Clark DA, Read JM (1998) Edaphic variation and the mesoscale distribution of tree species in a
neotropical rain forest. J For Ecol 86:101–112. doi:10.1046/j.1365-2745.1998.00238.x
Condit R, Engelbrecht BMJ, Pino D, Pérez R, Turner BL (2013) Species distributions in response to
individual soil nutrients and seasonal drought across a community of tropical trees. Proc Natl Acad Sci
USA 110(13):5064–5068. doi:10.1073/pnas.1218042110
Curtis JT, McInstosh RP (1951) An upland forest continuum in the prairie-forest border region of Wis-
consin. Ecology 32:476–496
de Guzman E, Umali RM, Sotalbo ED (1986) Guide to Philippine Flora and Fauna Vol III; Dipterocarps,
Non-Dipterocarps. NRMC, MNR and University of the Philippines, Los Banos
DAO 2007-01. Department of Environment and Natural Resources Administrative Order S. 2007-01. The
national list of threatened Philippine plants and their categories and the list of other wildlife species.
Forest Management Bureau, Philippines
Dimalanta CB, Suerte LO, Yumul JP Jr et al (2006) A Cretaceous supra-subduction oceanic basin source for
Central Philippine ophiolitic basement complexes: geological and geophysical constraints. Geosci J
10(3):305–320
Estrada-Medina H, Santiago LS, Graham RC et al (2013) Source water, phenology and growth of two
tropical dry forest tree species growing on shallow karst soils. Trees 27:1297–1307. doi:10.1007/
s00468-013-0878-9
Fernando ES, Suh MH, Lee J, Lee DK (2008) Forest formations of the Philippines. ASEAN Korea Envi-
ronmental Cooperation Unit (AKECU), National Seoul University, Seoul
Fernando ES, Bande MJM, Piollo RA et al (2009) Dipterocarpaceae of Bohol Island Philippines. Asia Life
Sci 18(1):121–138
Goldberg DE (1985) Effects of soil pH, competition, and seed predation on the distributions of two tree
species. Ecology 66:503–511
Guéze M, Paneque-Gálvez J, Luz AC et al (2013) Determinants of tree species turnover in a southern
Amazonian rain forest. J Veg Sci 24:284–295. doi:10.1111/j.1654-1103.2012.01461.x
Hamann A, Barbon EB, Curio E, Madulid DA (1999) A botanical inventory of a submontane tropical
rainforest on Negros Island, Philippines. Biodivers Conserv 8:1017–1031
Hijmans RJ, Cameron SE, Parra JL et al (2005) Very high resolution interpolated climate surfaces for global
land areas. Int J Climatol 25:1965–1978. doi:10.1002/joc.1276
Hubbell SP (2013) Tropical rain forest conservation and the twin challenges of diversity and rarity. Ecol
Evol 3(10):3263–3274. doi:10.1002/ece3.705
ISRIC (1995) Procedures for Soil Analysis. International Soil Reference and Information Center. (L. P. van
Reeuwijk, Ed.) Wageningen, The Netherlands
IUCN (2013) International Union for Conservation of Nature Red List Version 2013.1. www.iucnredlist.org.
Accessed 11 July 2013
Jahn R, Asio VB (1998) Soils of the tropical forests of Leyte, Philippines I: weathering, soil characteristics,
classification, and site qualities. In: Schulte A, Ruhiyat D (eds) Soils of tropical forest ecosystems.
Springer-Verlag, Berlin Heidelberg
Kaye TN, Meinke RJ, Kagan J et al (1997) Patterns of rarity in the Oregon Flora: implications for con-
servation and management. In: Kaye TN, Liston A, Love RM, Luoma DL, Meinke RJ, Wilson MV
(eds) Conservation and management of native plants and fungi. Native Plant Society of Oregon,
Corvallis
Kenfack D, Thomas DW, Chuyong G, Condit R (2007) Rarity and abundance in a diverse African forest.
Biodivers Conserv 16:2045–2074. doi:10.1007/s10531-006-9065-2

123
Biodivers Conserv (2014) 23:2183–2200 2199

Langenberger G, Martin K, Sauerborn J (2006) Vascular species inventory of a Philippine lowland rain
forest and its conservation value. Biodivers Conserv 15:1271–1301. doi:10.1007/s10531-005-2576-4
Lemmens RHMJ, Soerianegara I, Wong WC (eds) (1995) Plant resources of South-East Asia No. 5(2)
timber trees: minor commercial timbers. Backhuys Publishers, Leiden
Lucey JM, Tawatao N, Senior MJM et al (2014) Tropical forest fragments contribute to species richness in
adjacent oil palm plantations. Biol Conser 169:268–276. doi:10.1016/j.biocon.2013.11.014
Lunar BC, Arcega CVS (2011) Status of plant biodiversity in Mt. Malarayat, Batangas, Philippines. Asian J
Biodivers 90(2):1–13. doi:10.7828/ajob.v2i1.90
Lynagh F, Urich PB (2002) A critical review of buffer Zone theory and practice: a Philippine case study.
Soc Nat Res 15:129–145
Mangaoang EO, Pasa AE (2003) Preferred native tree species for smallholder forestry in leyte. Ann Trop
Res 25(1):25–30
McCune B, Mefford MJ (2006) PC-ORD multivariate analysis of ecological data, version 5.12. MjM
Software, Gleneden Beach
Milan PP, Margraf J (1994) Rainforestation farming: an alternative to conventional concepts. Ann Trop Res
16(4):17–27
Navarrete IA, Asio VB, Jahn R, Tatsuki K (2007) Characteristics and genesis of two highly weathered soils
in Samar, Philippines. Aust J Soil Res 45:153–163. doi:10.1071/SR06103
Newman MF, Burgess PF, Whitmore TC (1996) Manual of dipterocarps for foresters. Royal Botanic Garden
Edinburgh, UK. Center for International Forestry Research, Bogor
Nguyen H, Herbohn J, Firn J, Lamb D (2012) Biodiversity–productivity relationships in small-scale mixed-
species plantations using native species in Leyte province, Philippines. For Ecol Manag 274:81–90.
doi:10.1016/j.foreco.2012.02.022
Pennington RT, Prado DE, Pendry CA (2000) Neotropical seasonally dry forests and Quaternary vegetation
changes. J Biogeogr 27:261–273
Petit B, Montagnini F (2006) Growth in pure and mixed plantations of tree species used in reforesting rural
areas of the humid region of Costa Rica, Central America. For Ecol Manag 233:338–343. doi:10.1016/
j.foreco.2006.05.030
Potts MD, Ashton PS, Kaufman LS, Plotkin JB (2002) Habitat patterns in tropical rainforests: a comparison
of 105 plots in Northwest Borneo. Ecology 83:2782–2797
Rabinowitz D (1981) Seven forms of rarity. In: Synge H (ed) The biological aspects of rare plant con-
servation. Wiley, New York
Ranta P, Blom T, Niemela J et al (1998) The fragmented Atlantic rain forest of Brazil: size, shape and
distribution of forest fragments. Biodivers Conserv 7:385–403
Sahu SC, Dhal NK, Lal B, Mohanty RC (2012) Differences in tree species diversity and soil nutrient status
in a tropical sacred forest ecosystem on Niyamgiri hill range, Eastern Gahts, India. J Mt Sci 9:492–500.
doi:10.1007/s11629-012-2302-0
Saw LG, LaFrankie JV, Kochummen KM, Yap SK (1990) Fruit trees in a Malaysian rain forest. Econ Bot
45(1):120–136
Schade J, Gulmatico CV, Bal ML et al (1987) Forest resources of Region 8. Philippine-German forest
resources inventory project, Forest Management Bureau, DENR, Manila
Schneider T, Ashton MS, Montagnini F, Milan PP (2014) Growth performance of sixty tree species in
smallholder reforestation trials on Leyte, Philippines. N For 45:83–96. doi:10.1007/s11056-013-9393-5
Seamon JO, Mann SS, Steele OC, Utzurrum RCB (2006) Conservation value of remnant forest patches: tree
composition, spatial patterns, and recruitment in the Ottoville lowland forest American Samoa. Pac Sci
60(3):319–332. doi:10.1353/psc.2006.0018
ter Steege H, Pitman NCA, Sabatier D et al (2013) Hyperdominance in the Amazonian tree flora. Science
342:6156. doi:10.1126/science.1243092
Thompson SK (1992) Sampling. John Wiley and Sons, New York
Travaglia C, Baes AF, Tomas LM (1978) Geology of Samar island. soil and land resources appraisal project,
PHI/74/003. Bureau of soils, Department of Agriculture, United Nations Development Programme,
Food and Agriculture Organization of the United Nations. Field Document 1 (7):6
Turner CS, Tamblyn A, Dray R et al. (2003) The biodiversity of the Upper Imbang-Caliban Watershed,
North Negros Forest Reserve, Negros Occidental, Philippines. Coral Cay Conservation, London. www.
coralcay.org. Accessed 13 Oct 2013
Valencia R, Foster RB, Villa G et al (2004) Tree species distributions and local habitat variation in the
Amazon: large forest plot in eastern Ecuador. J Ecol 92:214–229. doi:10.1111/j.0022-0477.2004.
00876.x
Villarin RA (2013) Genetic variation patterns of Shorea contorta and Parashorea malaanonan (Diptero-
carpaceae) in the Philippines. PhD Dissertation. Georg-August-Universität Göttingen, Germany

123
2200 Biodivers Conserv (2014) 23:2183–2200

Whitford HN (1911) The forests of the Philippines. Part I: forest types and products. Department of the
Interior Bureau of Forestry, Bulletin No. 10, Manila
Wolda H (1981) Similarity indices, sample size and diversity. OecoIogia 50:296–302

123