Available via license: CC BY 4.0
Content may be subject to copyright.
Citation: Besi, E.E.; Mustafa, M.;
Yong, C.S.Y.; Go, R. Habitat Ecology,
Structure Influence Diversity, and
Host-Species Associations of Wild
Orchids in Undisturbed and
Disturbed Forests in Peninsular
Malaysia. Forests 2023,14, 544.
https://doi.org/10.3390/f14030544
Academic Editor: Claudia Mattioni
Received: 19 January 2023
Revised: 27 February 2023
Accepted: 6 March 2023
Published: 9 March 2023
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Habitat Ecology, Structure Influence Diversity, and
Host-Species Associations of Wild Orchids in Undisturbed
and Disturbed Forests in Peninsular Malaysia
Edward Entalai Besi * , Muskhazli Mustafa, Christina Seok Yien Yong and Rusea Go *
Department of Biology, Faculty of Science, Universiti Putra Malaysia, Serdang 43400, Selangor, Malaysia;
muskhazli@upm.edu.my (M.M.); chrisyong@upm.edu.my (C.S.Y.Y.)
*Correspondence: edwardentalai@upm.edu.my (E.E.B.); rusea@upm.edu.my (R.G.)
Abstract:
As an attempt to examine the causes of forest disturbance and degradation of the orchid
community, a comparative study on diversity and ecology in eight undisturbed and ten disturbed
forests in Peninsular Malaysia was conducted that varied in areas, elevations, vegetation types, and
disturbance regimes. Density and individual-based rarefaction curves were used to describe the
abundance. Univariate and multivariate analyses were also performed to explore the associations of
species abundance with biotic and abiotic factors. The study reported 239 orchid species belonging to
65 genera. Species richness, abundance, density, and diversity of orchids varied by locality. Higher
density of orchids (2.433 plants/km
2
) occurred in the undisturbed forests than in the disturbed forests
(0.228 plants/km
2
). As with the character of undisturbed forests, the temperature was between
27.8
±
0.3
◦
C and 31.2
±
0.2
◦
C, humid (77.1
±
1.2%–89.6
±
0.9%), and with low light intensity
(23.8
±
3.3
µ
mol m
−2
s
−1
–171.7
±
18.8
µ
mol m
−2
s
−1
), thus supporting the high density of the plants.
Disturbed forests had higher diversity (H = 4.934 and 1-D = 0.990) and abundance (183 species of
57 genera) but were determined to be highly influenced by the higher abundance of epiphytic orchids
on the fallen trees and ease of accessibility in the logged forests. Terrestrial and mycoheterotroph
orchids were much lower in density and abundance in the disturbed habitat indicating a gradual
reduction in their niche availability following the disturbance. Additionally, the ecology data show
that the microclimate conditions of the canopy-covered forest was influenced by proximity to the
logged area which had eventually reduced the orchids’ habitat quality. Furthermore, the results show
that the abundance of epiphytic orchid communities was associated with the host plant characteristics.
Host types and bark texture preference were apparent for the epiphytic orchid species, with certain
types and textures hosting more orchid species than others. Overall results show that extreme
temperature, humidity, and light intensity caused by the canopy opening inflicted damages to the
habitat conditions and bark textures of the host plants and limits recolonisation of the orchids in
the disturbed forests. The species diversity and density patterns of orchids in undisturbed and
disturbed forests revealed in this study provide a baseline for conservationists, policy makers, and
forest authorities in expanding the understanding of the forest ecology and vegetation along the
disturbance gradient, forest regeneration, and criteria for plant selection for forest restoration in
Peninsular Malaysia.
Keywords: diversity; environmental factor; epiphytic distribution; host plant; logging; orchid
1. Introduction
Orchids are diverse and a very good group for studying diversity, ecology, distribution,
and host-species associations in undisturbed and disturbed habitats [
1
]. The diversity
study focused on the orchids occurring in Peninsular Malaysia, where previously a total of
982 orchid species were recorded and most of them were epiphytes (personal observation).
Apart from focusing on their diversity, ecology, distribution patterns, and conservation,
Forests 2023,14, 544. https://doi.org/10.3390/f14030544 https://www.mdpi.com/journal/forests
Forests 2023,14, 544 2 of 30
this paper explores the associations between diversity and distribution of epiphytic orchids
and host characteristics in different vegetation types of Peninsular Malaysia. Although the
association between epiphyte diversity and different host characteristics are well studied for
plants, very little is known about variation in the abundance of epiphytic orchids and their
host associations in tropical rainforests. Host plant characteristics play an important role in
colonisation of host-dependent species [
2
–
4
]. In studies of relationship between tree size
and epiphyte species richness, the host’s size matters due to greater bark surface available
for colonisation on large trees and the creation of additional microhabitats [
5
,
6
]. Host
position and type, and growth of epiphytic orchid species are influenced by microclimate
conditions, including annual precipitation, light intensity, and humidity [
3
,
7
]. Conservation-
wise, the degree of orchids’ host specificity is important to look at, because being host
specialists, wild orchids are threatened by co-extinction with their hosts due to habitat
alterations and climate change [8–10].
Lack of monitoring programs assessing the before or aftereffects of disturbances in
Malaysian rainforests is one of the major limitations to quantifying the commonalities
and differences of disturbance effects on forest ecosystem and its inhabitants. Epiphytes
are a diverse group with very interesting niche ecology [
11
–
13
]. Unlike terrestrial species,
epiphytes are highly reliant on their hosts and are thus likely to be affected by host species
and host traits [
14
–
17
]. Notwithstanding their economical and conservation importance, the
understanding of the wild orchid diversity and ecology under wide range of microclimate
conditions in the undisturbed and disturbed forests is still rather poor, especially when
the impacts come under different disturbance regimes. A comparative study of different
undisturbed and disturbed forest types including scoring for different agents of disturbance
will give unbiased and general views on the effect of forest and ecological disturbance on
the orchid community. Additionally, if only protected areas are targeted, more wild orchids
will be lost, and these species could disappear from large areas of former range. Thus,
the present study includes several forest types, vegetation types, and disturbance regimes.
This diversity and ecology study determines the suitability of the plants, especially trees as
protectors and homes to epiphytic plants along with threats and limiting factors.
Specifically, in this paper, the following questions were addressed: (1) Does the di-
versity of orchids in undisturbed and disturbed forests of Peninsular Malaysia differ?
(2) What are the factors determining diversity and composition of orchids in the undis-
turbed and disturbed forests? (3) What is the significance of ecological alterations on the
orchid community in two adjacent forests that are located next to each other or connected
to each other, but with distinct levels of disturbance? (4) How much do the number of host
plants and host characteristics influence the number of epiphytic orchid species?
2. Materials and Methods
2.1. Study Area
Samplings were conducted between November 2016 and April 2021 in 18 different
localities in Peninsular Malaysia (Figure 1). The systematic sampling involved choosing
exploration sites in Peninsular Malaysia. We collected information on each of the site con-
ditions (e.g., area, vegetation type, elevation, disturbance regime, and disturbance origin)
(Tables 1and 2). Major disturbances were classified as either natural or anthropogenic in
origin. To disentangle the effects of anthropogenic and natural disturbance regimes, there
were two main forest types varying in vegetation: undisturbed (Table 1) and disturbed
(Table 2) forests, with several subsidiary study sites, eight undisturbed and ten disturbed
forests, were selected. The sub-classified disturbed forest types were defined according
to the vegetation, disturbance, and period of disturbance. Classification of the vegetation
types was based on [
18
] (modified from [
19
], and [
20
] for Bukit Pedu, Kedah. These sites
represent some of the major ecosystem and vegetation types in Peninsular Malaysia. To
characterise and compare disturbance regimes amongst disturbed forest types, the Dis-
turbance Index (DI) based on five variables (Severity (D), Extension (E), Frequency (F),
Duration (D), and Patch Type (PT)) were evaluated following a scoring method and ordinal
Forests 2023,14, 544 3 of 30
variables modified from [
21
] (Table 3). The DI was calculated by multiplying the assigned
scores describing the five disturbance variables: DI = S
×
E
×
F
×
D
×
P
×
T. A low DI
indicates a low impact of a given disturbance agent for a particular ecosystem, whereas a
high DI indicates a large impact.
Forests 2023, 14, x FOR PEER REVIEW 4 of 32
Figure 1. Map showing the study localities (modified and reproduced with permission from JPSM
(2019)): 1: Hulu Setiu, logged lowland dipterocarp forest (LHS), 2: Bukit Batu Kapal, lowland ripar-
ian forest (BK), 3: Gunung Perlis, hill dipterocarp forest (GP), 4: Bukit Rongkit, hill limestone forest
(BR), 5: Bukit Batu Kapal, forest margin of a mixed dipterocarp forest (BK(M)), 6: Bukit Batu Kapal,
logged hill dipterocarp forest (LBK), 7: Bukit Sekayu, lowland riparian forest (BS), 8: Padang 7, hill
mixed dipterocarp forest (PG7), 9: Taman Rimba Komanwel, lowland dipterocarp forest (TRK), 10:
Lata Kekabu, lowland riparian forest (KB), 11: Lata Lawin, lowland riparian forest (LW), 12: Lata
Figure 1.
Map showing the study localities (modified and reproduced with permission from JPSM
(2019)): 1: Hulu Setiu, logged lowland dipterocarp forest (LHS), 2: Bukit Batu Kapal, lowland riparian
Forests 2023,14, 544 4 of 30
forest (BK), 3: Gunung Perlis, hill dipterocarp forest (GP), 4: Bukit Rongkit, hill limestone forest
(BR), 5: Bukit Batu Kapal, forest margin of a mixed dipterocarp forest (BK(M)), 6: Bukit Batu Kapal,
logged hill dipterocarp forest (LBK), 7: Bukit Sekayu, lowland riparian forest (BS), 8: Padang 7,
hill mixed dipterocarp forest (PG7), 9: Taman Rimba Komanwel, lowland dipterocarp forest (TRK),
10: Lata Kekabu, lowland riparian forest (KB), 11: Lata Lawin, lowland riparian forest (LW), 12: Lata
Tanjung Kala, lowland riparian forest (TK), 13: Bintang Hijau, logged lowland dipterocarp forest
(LB), 14: Bukit Pedu, conglomerate hill forest (BP), 15: Gawi, logged lowland dipterocarp forest (LG),
16: Petuang, logged lowland dipterocarp forest (LP), 17: Kuala Koh, lowland dipterocarp forest (KH),
18: Tanah Merah, lowland forest (TM). Source: Jabatan Perhutanan Semenanjung Malaysia (JPSM).
Table 1.
General characteristics of undisturbed forests in Peninsular Malaysia considered in this study.
Sites States Vegetation Types Elevation
(m)
No. of
Transect
Plots
Totals Transect
Size/Area (km2)
Padang 7 (Rompin) (PG7) Pahang Hill mixed dipterocarp forest (with
coastal mountain outcrops or kerangas)268–600 4 0.9
Bukit Rongkit (BR) Perlis Hill limestone forest 300–400 2 14.4
Taman Rimba
Komanwel (TRK) Selangor Lowland dipterocarp forest 100–200 6 16.8
Bukit Batu Kapal (BK)
Terengganu
Lowland riparian forest 100–200 5 5.6
Lata Kekabu (KB) Perak Lowland riparian forest 100–200 5 5.5
Lata Lawin (LW) Perak Lowland riparian forest 100–200 3 3.6
Lata Tanjung Kala (TK) Perak Lowland riparian forest 100–200 2 12.0
Bukit Pedu (BP) Kedah Conglomerate hill forest 200–400 1 9.0
Total 28 67.8
Forests 2023,14, 544 5 of 30
Table 2. General characteristics of disturbed forests in Peninsular Malaysia considered in this study.
Site State Vegetation Types S E F D PT DI Disturbance
Regime Disturbance Origin Elevation
(m)
No. of
Transect
Plots
Totals
Transect
Size/Area
(km2)
Bukit Sekayu (BS)
Terengganu
Lowland riparian forest 1 1 1 3 1 3 Low
Fragmentation
caused by palm
oil estate
90–100 1 11.8
Gunung Perlis (GP) Perlis Hill dipterocarp
forest (granite) 1 1 1 3 1 3 Low
Erosion of trail
due to heavy
ecotourism activities
500–800 4 34.0
Bukit Batu Kapal [BK(M)]
Terengganu
Mixed dipterocarp forest 1 1 1 1 1 1 Low
Timber extraction
(a forest margin
bordered by
logged forests)
100–400–
600 1 12.6
Tanah Merah (TM) Kelantan Lowland forest 1 1 1 1 1 1 Low Timber extraction,
plantation, mud-flood
40–100 2 0.3
Hulu Setiu (LHS)
Terengganu
Lowland dipterocarp forest
2 2 1 2 1 8 Mild Timber extraction 40–100 6 13.0
Kuala Koh (KH) Kelantan
Lowland dipterocarp forest
2 2 1 2 1 8 Mild
Timber extraction
(an encroachment),
mud-flood
90–130 8 0.8
Bukit Batu Kapal (LBK)
Terengganu
Hill dipterocarp forest 3 2 2 2 1 24 Severe Timber extraction 300–500 1 33.3
Bintang Hijau (LB) Perak
Lowland dipterocarp forest
3 2 3 3 2 108 Severe Timber extraction 100–200 3 1200.0
Gawi (LG)
Terengganu
Lowland dipterocarp forest
3 2 3 3 2 108 Severe Timber extraction 90–250 1 19.5
Petuang (LP)
Terengganu
Lowland dipterocarp forest
3 2 3 3 2 108 Severe Timber extraction 200–330 2 53.0
Total 29 1409.7
Notes: Disturbance Index (DI) = Severity (D)
×
Extension (E)
×
Frequency (F)
×
Duration (D)
×
Patch Type (PT). Interpretation modified from Calderon-Aguilera et al. (2012): Low DI
(1–5) = low impact, Intermediate DI (6–20) = mild impact, High DI (>20) = large impact.
Forests 2023,14, 544 6 of 30
Table 3.
Ordinal variables assessed to characterise and compare disturbance regimes amongst the
disturbed studied sites.
Variables Definition Scoring Value
Severity (S) Loss in ecosystem quality based on the
different types of disturbance
(1) small (disturbed ground vegetation due to flood or heavy
human trampling)
(2) medium (disturbed ground vegetation and forest canopy due
to fallen trees and mud flood)
(3) large (massive removal of ground vegetation and forest
canopy due to clear-felled timber extraction and logs’ dragging
extraction by tractor
Extension (E) Area affected by a given disturbance
(1) small (>20% of area)
(2) medium (>20% ≤50% of area)
(3) large (>50% of area)
Frequency (F) Occurrence of disturbance event in a
given time period
(1) low (occurring at intervals longer than 10 years)
(2) medium (once every 2–10 years)
(3) high (once every year or seasonal)
Duration (D) Length in time of the disturbance event
(1) <1 year
(2) <3 years
(3) >3 years
Patch type (PT) Type of habitat patch created by
the disturbance
(1) embedded (disturbed patches within a continuous
undisturbed ecosystem)
(2) isolated (disturbed patches isolated from an
undisturbed ecosystem)
2.2. Sampling Design, Samples, and Data Collection
Samplings were carried along the existing and off-road four-wheel trails using the line-
transect method designed at accessible trail(s) in the selected study sites and established
randomly by selecting a point along the trail in each locality. To accommodate sampling
in the large areas such as logging sites and skid trails, the study areas were calculated
based on the distance travelled of each trail multiplied by the width of 5–10 metres on each
side of the trail, as shown in Figure 2. Observers walk along lines and count all orchid
plants observed within the specified width. The length, width, and number of transect
plots were not fixed considering the topography, accessibility, time constraint, movement
restrictions due to the pandemic, and budget. Due to the limitations, the comparative study
focused on different forest types irrespective of the total size area. Square kilometre (km
2
)
is used as the unit for total transect plots’ area for ease of density calculation, graph scaling,
and plotting.
One individual of the flowering specimens was collected and all occurrence pertaining
to the same species were successively pooled for each transect and trail. Samples with
floral structures were preserved using a standard herbarium technique [
22
] and voucher
specimens were deposited in the Herbarium of Universiti Putra Malaysia (UPM) at the
Department of Biology, Faculty of Science. The robust non-flowering individuals were
brought back to UPM and transplanted into the ex situ conservatory as living collections.
However, some of the non-flowering individuals that are also fragile, minute size, and
found growing in few populations were not collected during the first visit until flowering.
The flowering individuals were then collected during the revisits. Additionally, some sterile
individuals were simply determined at generic level or to the closest affinity on the basis
of their vegetative morphological characters. The exact locality is withheld to protect the
population from illegal collections.
Forests 2023,14, 544 7 of 30
Forests 2023, 14, x FOR PEER REVIEW 8 of 32
Figure 2. An example of a line-transect (1000 m × 20 m transect plot). The total transect plots’ area
was converted to the unit of square kilometre (km2).
One individual of the flowering specimens was collected and all occurrence pertain-
ing to the same species were successively pooled for each transect and trail. Samples with
floral structures were preserved using a standard herbarium technique [22] and voucher
specimens were deposited in the Herbarium of Universiti Putra Malaysia (UPM) at the
Department of Biology, Faculty of Science. The robust non-flowering individuals were
brought back to UPM and transplanted into the ex situ conservatory as living collections.
However, some of the non-flowering individuals that are also fragile, minute size, and
found growing in few populations were not collected during the first visit until flowering.
The flowering individuals were then collected during the revisits. Additionally, some ster-
ile individuals were simply determined at generic level or to the closest affinity on the
basis of their vegetative morphological characters. The exact locality is withheld to protect
the population from illegal collections.
For species identification and evaluation of each species’ distribution status, reliable
taxonomic and floristic literatures were referred and studied, e.g., [23–26]. Digitalised im-
ages of herbarium collections, botanical drawings, and records deposited in the National
Herbarium of the Netherlands (NHN) accessed through Browse Dutch Natural History
Collections: BioPortal (Naturalis) (http://bioportal.naturalis.nl/ (accessed on 27 October
2022)), Herbarium of Singapore Botanic Gardens (SING) accessed through BRAHMS
Online managed by the University of Oxford (http://herbaria.plants.ox.ac.uk/bol/sing (ac-
cessed on 27 October 2022)), Swiss Orchid Foundation [https://orchid.unibas.ch/in-
dex.php/en/ (accessed on 27 October 2022)), Kew Herbarium Catalogue
(http://apps.kew.org/herbcat/gotoSearchPage.do (accessed on 27 October 2022)), Natural
History Museum Specimen Collection (https://data.nhm.ac.uk/ (accessed on 27 October
2022)), Herbarium of Aarhus University (AAU)(https://www.aubot.dk/search_form.php
(accessed on 27 October 2022)), Museum National D’Histoire Naturelle (MNHN)
(https://science.mnhn.fr/all/search (accessed on 27 October 2022)), and Plants of the World
Online (POWO) (http://www.plantsoftheworldonline.org/ (accessed on 27 October 2022))
were examined prior to the taxonomic treatment and assessment on the range of distribu-
tion for each species. The accepted names were validated via KEW World Checklist of
Selected Plant Families (WCSP) [27] and POWO.
Occurrence of the host plants containing epiphytic orchids was recorded within the
studied transect plots. For each sampled host plant, parameters, and characteristics:
height, bark’s texture, canopy’s structure, crown’s forms and classes, and substrate cover
were recorded in a spreadsheet. Height was measured either using clinometer and visual
Figure 2.
An example of a line-transect (1000 m
×
20 m transect plot). The total transect plots’ area
was converted to the unit of square kilometre (km2).
For species identification and evaluation of each species’ distribution status, reliable
taxonomic and floristic literatures were referred and studied, e.g., [
23
–
26
]. Digitalised
images of herbarium collections, botanical drawings, and records deposited in the National
Herbarium of the Netherlands (NHN) accessed through Browse Dutch Natural History
Collections: BioPortal (Naturalis) (http://bioportal.naturalis.nl/ (accessed on 27 Octo-
ber 2022)), Herbarium of Singapore Botanic Gardens (SING) accessed through BRAHMS
Online managed by the University of Oxford (http://herbaria.plants.ox.ac.uk/bol/sing
(accessed on 27 October 2022)), Swiss Orchid Foundation [https://orchid.unibas.ch/index.
php/en/ (accessed on 27 October 2022)), Kew Herbarium Catalogue (http://apps.kew.
org/herbcat/gotoSearchPage.do (accessed on 27 October 2022)), Natural History Museum
Specimen Collection (https://data.nhm.ac.uk/ (accessed on 27 October 2022)), Herbar-
ium of Aarhus University (AAU)(https://www.aubot.dk/search_form.php (accessed on
27 October 2022)), Museum National D’Histoire Naturelle (MNHN) (https://science.mnhn.
fr/all/search (accessed on 27 October 2022)), and Plants of the World Online (POWO)
(http://www.plantsoftheworldonline.org/ (accessed on 27 October 2022)) were examined
prior to the taxonomic treatment and assessment on the range of distribution for each
species. The accepted names were validated via KEW World Checklist of Selected Plant
Families (WCSP) [27] and POWO.
Occurrence of the host plants containing epiphytic orchids was recorded within the
studied transect plots. For each sampled host plant, parameters, and characteristics: height,
bark’s texture, canopy’s structure, crown’s forms and classes, and substrate cover were
recorded in a spreadsheet. Height was measured either using clinometer and visual
interpretation or measuring tape for fallen trees. Each character was described and pho-
tographed on the field. The presence of epiphytic orchids was recorded for the entire
host plants. The host species and associated epiphytic orchids were mostly identified in
the field and verified using different local floras, e.g., [
28
–
38
]. The unidentified species
of host plants were photographed in detail along with their flowers, leaves, and fruits
(if present) and later identified with expert consultation. Studying epiphyte diversity is
challenging because it is difficult to observe and identify the plants in the tree canopy from
the ground [
39
]. Each host plant was examined from the ground at different points, thereby
assuring a clear view of all the orchids and host’s parts [
4
,
40
]. Hence, for large and tall host
plants, the inaccessible epiphyte growing up on it were only photographed and captured
using a Nikon D5100 and a macro lens AF-S DX NIKKOR 55–300 mm f/4.5–5.6 G ED VR
Forests 2023,14, 544 8 of 30
(very high-powered telephoto zoom) as evidence, and on-field visual identification were
employed.
2.3. Ecological Study Design and In Situ Measurement of Microclimate Data
In this study, intact undisturbed area and adjacent disturbed area were compared
to understand the significant impacts of ecological dynamics towards orchid diversity.
Four ecological parameters were logged from morning (9 am) to afternoon time (4 pm)
under the influence of varying spatial-temporal conditions; temperature (
◦
C), relative air
humidity (10%–100%), and light intensity (
µ
mol m
−2
s
−1
) (number of photons received per
unit time (s) on a unit area (m)) of the study areas, and bark moisture (0%–100%) of the host
plants. The temperature and humidity were measured with Extech Hygro-Thermometer
(Extech Instruments, Inc., Nashua, NH, USA) at 2 m above the ground during the study
period [
41
]. The light intensity was first measured in Lux unit with Milwaukee MW700
Standard Portable Lux Meter (Milwaukee Instruments, Inc., Rocky Mount, NC, USA) and
then value was converted to the Photosynthetic Photon Flux Density (PPFD) unit. Given an
illuminance value (Lux), similarly, we calculated the PPFD in micromoles per second per
square metre (
µ
mol m
−2
s
−1
) for the given light source [
42
,
43
]. Unit Lux is simply based on
visual sensitivity and does not provide information on the energy or photon content of light,
which truly influence the photosynthesis or sugar production in the leaf. Hence, in order to
understand better the light intensity for a study relating to plant responses, the suitable
unit is the
µ
mol m
−2
s
−1
. The bark moisture was measured with HoldPeak HP-2GD Wood
Moisture Meter (Zhuhai Jida Huapu Instrument Co., Ltd., Zhuhai, Guangdong Province,
China) at 2 m above the basal area.
2.4. Diversity Analyses
Recorded information about the total individual species were analysed statistically and
mathematically for clearer identification of diversity in the area. Diversity analyses were
estimated using PAleontological STatistics Version 4.04 (PAST4) [
44
,
45
]. To determine the
species richness and species evenness, Shannon–Wiener Diversity Index (H) and Simpson’s
Diversity Index (1-D) were used in this study. Inclusion of both diversity indices improves
the output information of the dataset, which is unique for each community or sample
analysed. The H and 1-D indices allow data on species richness and relative abundance
to be combined [
46
]. The Simpson Index was explained by presence of the predominant
species, and the Shannon Index assumed that individuals were randomly selected, and
that all species were represented in the sample [
47
]. The effect of the sample size is
generally insignificant for both indices. However, these methods could not tell which
factors contributed more to the value. Hence, Evenness (E) was used to determine how
close in numbers each species in each studied area were [
48
]. Higher values indicate greater
community diversity and more even distribution of different individuals within a species.
Dominance (D) gives the abundance of the most abundant species and ranges from 0 (all
taxa are equally present) to 1 (one taxon dominates the community completely) [44].
2.5. Data Analyses
Abundance was calculated as Relative abundance of orchid (%Ao): [Number of clump
or population of a particular orchid species within the transect plots
÷
Total number of
all orchid clumps or populations of the transect plots]
×
100. When sampling design is
transect-based, orchid and host plants are sampled in unequal numbers due to the variation
in size of the studied areas or trails. Hence, the following parameters were computed;
Density of the orchid plants within a transect plot (plants/km
2
): Number of orchids plants
within the transect plot (plants)
÷
Total area of the transect plots (km
2
). An ordination
method, clustering with UPGMA (Unweighted Pair Group Method with Arithmetic Mean)
and Q-mode clustering (grouping variables or associations) by transposing the data ma-
trix [
45
], was used to look at similarity of host plants (bark textures) amongst the studied
vegetation alliances. The similarity was tested using Bray–Curtis Similarity Coefficient [
49
].
Forests 2023,14, 544 9 of 30
Bray–Curtis is a popular similarity index for abundance data (Hammer et al., 2001). The or-
dination method was carried out using PAleontological STatistics Version 4.04 (PAST4) [
45
].
2.6. Statistical Analyses
Ecological data obtained were analysed using software package IBM SPSS version
24 (IBM Corp., Chicago, IL, USA). The measured microclimate parameters were sub-
jected to Shapiro–Wilk for a normality distribution test. As the parametric assump-
tions could not be obtained for all dependant variables even through transformation
(Supplementary Table S1a,c,e,g), non-parametric tests were used for all analyses. The
Mann–Whitney test was used to determine the significant relationship of the parameters
between the undisturbed and disturbed forests. Statistical significance was determined at
p< 0.05 (Supplementary Table S1b,d,f). Following the study [
50
], the significant difference
of mean between groups were determined by comparing directly between the minimum
values recorded for undisturbed forests with minimum values recorded for disturbed
forests and evaluated accordingly for the maximum values. Kruskal–Wallis one-way analy-
sis of variance (ANOVA) was used to determine the significant difference in daily hours’
ecological parameters between undisturbed, disturbed, and disturbed ‘non-logging sites’
forests with distinctively different disturbance types (Supplementary Tables S1h,j,k). The
post hoc for Kruskal–Wallis was conducted using the Mann–Whitney test with multiple
pairwise comparisons (Supplementary Table S1i,l). To ensure that Type I error does not t
exceed 0.05, Bonferroni correction was used by dividing
α
of 0.05 by the number of tests
conducted. Therefore, the statistical significance was determined at p< 0.017. Comparing
species richness among sites or samples is a statistical challenge because the observed
number of species is sensitive to the number of individuals counted or the area sam-
pled [
51
]. Hence, to describe the orchid species richness in the different localities, we used
a rarefaction function for each region. Rarefaction is a method for comparing the species
richness of samples of different sizes by calculating species richness for a given number
of individuals based on rarefaction curves, which are plots of the number of species as a
function of the number of samples [
52
], in our case number of clumps (one clump = one
individual = one occurrence). We conducted individual-based rarefaction as implemented
in EstimateS 9.1.0 [53].
2.7. Terminology
Following the study [
54
], the terms ‘host’ and ‘host species’ imply that the focal
epiphyte taxon has been observed on the plant individual or species. Since hosts are of
several growth forms (trees, shrubs, lianas), the term ‘host plants’ is used rather than ‘host
tree’. Note that, in unambiguous contexts, ‘host’ is often used as a short form of ‘host
species’ or ‘host plant’. Therefore, the unoccupied (by epiphytic orchids) plant species were
excluded from the analysis. In the current study, ‘undisturbed forest’ and ‘disturbed forest’
are the general classification used for the forest types. The terms ‘undisturbed forest’ and
‘disturbed forest’ were adopted from [
55
]. The synonymous terms such as undisturbed and
disturbed areas, or undisturbed and disturbed vegetation have been used in other works
of literature on biodiversity in the tropics or Malaysia, such as [
18
,
56
–
61
]. ‘Undisturbed
forest’ are primary or protected forests that have experienced little to no recent human
disturbance, whereas ‘disturbed forest’ are forest that have experienced disastrous and
large disturbance such as deforestation, extreme floods, tree falls, and human trampling.
3. Results
3.1. Species Richness and Abundance
In total, 389 orchid specimens confined to 239 species and 65 genera were collected
(Supplementary Table S2). A plot of species accumulation curve allows comparison of a
number of species at different levels of collecting effort. Using the number of individuals,
the disturbed forest was the most species-rich for equal sampling effort if compared to the
undisturbed forest (Figure 3). Additionally, the curve for the undisturbed forest stopped
Forests 2023,14, 544 10 of 30
at 250 individuals and appeared to approach an asymptote, whereas the curve for the
disturbed forest is longer but steeper, indicating that there may be more species found in
the disturbed forest if the sampling continues. This may indicate a higher abundance of
species which can be found in heterogeneous and larger area sites [
62
], but nothing of the
sort for logging areas as the orchid abundance is influenced by the number of fallen trees
and ease of collections for epiphytic orchids on the fallen trees along the skid trails [50].
Forests 2023, 14, x FOR PEER REVIEW 11 of 32
and ‘disturbed forest’ were adopted from [55]. The synonymous terms such as undis-
turbed and disturbed areas, or undisturbed and disturbed vegetation have been used in
other works of literature on biodiversity in the tropics or Malaysia, such as [18,56–61].
‘Undisturbed forest’ are primary or protected forests that have experienced little to no
recent human disturbance, whereas ‘disturbed forest’ are forest that have experienced dis-
astrous and large disturbance such as deforestation, extreme floods, tree falls, and human
trampling.
3. Results
3.1. Species Richness and Abundance
In total, 389 orchid specimens confined to 239 species and 65 genera were collected
(Supplementary Table S2). A plot of species accumulation curve allows comparison of a
number of species at different levels of collecting effort. Using the number of individuals,
the disturbed forest was the most species-rich for equal sampling effort if compared to the
undisturbed forest (Figure 3). Additionally, the curve for the undisturbed forest stopped
at 250 individuals and appeared to approach an asymptote, whereas the curve for the
disturbed forest is longer but steeper, indicating that there may be more species found in
the disturbed forest if the sampling continues. This may indicate a higher abundance of
species which can be found in heterogeneous and larger area sites [62], but nothing of the
sort for logging areas as the orchid abundance is influenced by the number of fallen trees
and ease of collections for epiphytic orchids on the fallen trees along the skid trails [50].
Figure 3. Orchid species accumulation curves (individual-based rarefaction) within grid points of
different land-use intensities.
Species richness in the disturbed forests was 183 species with the logging sites har-
bouring the most abundant orchids species if compared to the undisturbed forests with
100 species (Figure 4A). Despite that, undisturbed forests had a higher orchid density (2.4
orchids) when compared with the disturbed forests (0.2 orchids). However, disturbed
habitats, especially the logging sites, showed greater impoverishment than the undis-
turbed habitats as very few surviving plants were found or collected. Similarly, Agrosto-
phyllum stipulatum was recorded with highest relative abundance in both undisturbed
(3.6%) and disturbed forests (4.6%). The species composition of the orchids differed sig-
nificantly in between undisturbed and disturbed forests. Disturbed forests comprised
mostly epiphytic orchids (Figure 4B), of which the highest collection was from logging
Figure 3.
Orchid species accumulation curves (individual-based rarefaction) within grid points of
different land-use intensities.
Species richness in the disturbed forests was 183 species with the logging sites har-
bouring the most abundant orchids species if compared to the undisturbed forests with
100 species (Figure 4A). Despite that, undisturbed forests had a higher orchid density
(2.4 orchids) when compared with the disturbed forests (0.2 orchids). However, dis-
turbed habitats, especially the logging sites, showed greater impoverishment than the
undisturbed habitats as very few surviving plants were found or collected. Similarly,
Agrostophyllum stipulatum was recorded with highest relative abundance in both undis-
turbed (3.6%) and disturbed forests (4.6%). The species composition of the orchids differed
significantly in between undisturbed and disturbed forests. Disturbed forests comprised
mostly epiphytic orchids (Figure 4B), of which the highest collection was from logging
sites. Terrestrial species occurred the most in the undisturbed habitats. Only few to none
were marginally more frequent in a more open and disturbed habitat. Additionally, species
density of orchids based on growth habits was lower in disturbed habitats than in the undis-
turbed or pristine ones (Figure 4C). The two forest types were Dendrobium and Bulbophyllum
as of the most abundant genera which were also mostly epiphytic (Figure 5).
Forests 2023,14, 544 11 of 30
Forests 2023, 14, x FOR PEER REVIEW 12 of 32
sites. Terrestrial species occurred the most in the undisturbed habitats. Only few to none
were marginally more frequent in a more open and disturbed habitat. Additionally, spe-
cies density of orchids based on growth habits was lower in disturbed habitats than in the
undisturbed or pristine ones (Figure 4C). The two forest types were Dendrobium and
Bulbophyllum as of the most abundant genera which were also mostly epiphytic (Figure
5).
Figure 4. Comparison of species abundance and density between undisturbed and disturbed forests:
(A) Species and genera abundance, (B) Relative abundance based on growth habit, (C) Species den-
sity based on growth habit.
Figure 4.
Comparison of species abundance and density between undisturbed and disturbed forests:
(
A
) Species and genera abundance, (
B
) Relative abundance based on growth habit, (
C
) Species density
based on growth habit.
Forests 2023,14, 544 12 of 30
Forests 2023, 14, x FOR PEER REVIEW 13 of 32
Figure 5. Comparison of species abundance based on genera for (A) undisturbed and (B) disturbed
forests in Peninsular Malaysia.
3.2. Species Diversity
Shannon–Wiener Diversity Index (H) (Figure 6A) and Simpson’s Diversity Index (1-
D) (Figure 6B) were showing similar patterns of diversity for the undisturbed and dis-
turbed forests. Both indices showed that disturbed forests harboured higher species di-
versity than the undisturbed forests. Theoretically, H is strongly influenced by species
richness, evenness, and presence of rare species. Meanwhile, 1-D gives more weight to
evenness and presence of common species. In this sense, H has an advantage over 1-D
because it depends more on species richness and less on abundance, so it is very sensitive
to even small diversity changes, and thus is widely used to assess the actual state of envi-
ronment. Although the finding suggests that disturbed forests could retain a high richness
of orchid taxa, there are several detrimental aspects that need consideration—the high
abundance of orchids in the disturbed forests might overstate the species richness. The
Figure 5. Comparison of species abundance based on genera for (A) undisturbed and (B) disturbed
forests in Peninsular Malaysia.
3.2. Species Diversity
Shannon–Wiener Diversity Index (H) (Figure 6A) and Simpson’s Diversity Index
(1-D) (Figure 6B) were showing similar patterns of diversity for the undisturbed and
disturbed forests. Both indices showed that disturbed forests harboured higher species
diversity than the undisturbed forests. Theoretically, H is strongly influenced by species
richness, evenness, and presence of rare species. Meanwhile, 1-D gives more weight to
evenness and presence of common species. In this sense, H has an advantage over 1-D
because it depends more on species richness and less on abundance, so it is very sensitive
to even small diversity changes, and thus is widely used to assess the actual state of
environment. Although the finding suggests that disturbed forests could retain a high
richness of orchid taxa, there are several detrimental aspects that need consideration—the
high abundance of orchids in the disturbed forests might overstate the species richness.
Forests 2023,14, 544 13 of 30
The low orchid diversity in the undisturbed forest is supported by the high dominance
of the most dominant species (D = 0.015) (Figure 6C), and the slightly higher evenness
in the undisturbed forest (E = 0.844) does not seem to affect the H and 1-D values much
(Figure 6D). Additionally, the higher diversity of orchids in the disturbed forests might be
influenced by the higher number of rare species found on the fallen trees in the logged areas
(Figure S1). Moreover, forest margin had lower species diversity (Figure 7A) and density
(Figure 7C) than the adjacent undisturbed forest with no terrestrial orchid recorded in the
area. It suggests the altered environment caused by the adjacent logging activity reduced
the orchid species richness. In logging sites, orchid species were greatest in diversity if
compared to the undisturbed forests (Figure 7E), but lower in species density (Figure 7G).
It only comprised of epiphytic orchids (Figure 7H). Further details on the occurrence,
diversity, and density of orchids in each study site are presented in Table 4.
Forests 2023, 14, x FOR PEER REVIEW 14 of 32
low orchid diversity in the undisturbed forest is supported by the high dominance of the
most dominant species (D = 0.015) (Figure 6C), and the slightly higher evenness in the
undisturbed forest (E = 0.844) does not seem to affect the H and 1-D values much (Figure
6D). Additionally, the higher diversity of orchids in the disturbed forests might be influ-
enced by the higher number of rare species found on the fallen trees in the logged areas
(Figure S1). Moreover, forest margin had lower species diversity (Figure 7A) and density
(Figure 7C) than the adjacent undisturbed forest with no terrestrial orchid recorded in the
area. It suggests the altered environment caused by the adjacent logging activity reduced
the orchid species richness. In logging sites, orchid species were greatest in diversity if
compared to the undisturbed forests (Figure 7E), but lower in species density (Figure 7G).
It only comprised of epiphytic orchids (Figure 7H). Further details on the occurrence, di-
versity, and density of orchids in each study site are presented in Table 4.
Figure 6. Overall comparison of species diversity between the undisturbed and disturbed forests:
(A) Shannon (H), (B) Simpson (1-D), (C) Dominance, (D) Evenness (e^H/S).
Figure 6.
Overall comparison of species diversity between the undisturbed and disturbed forests:
(A) Shannon (H), (B) Simpson (1-D), (C) Dominance, (D) Evenness (eˆH/S).
Forests 2023,14, 544 14 of 30
Forests 2023, 14, x FOR PEER REVIEW 15 of 32
Figure 7. Comparison of species diversity and density for Adjacent Undisturbed vs. Disturbed For-
ests and Undisturbed Forests vs. Logging Sites: (A,E) Shannon (H), (B,E) Dominance (D), (C,G)
Density (plants/km2) based on forest types, (D,H) Density (plants/km2) based on growth habit.
Figure 7.
Comparison of species diversity and density for Adjacent Undisturbed vs. Disturbed
Forests and Undisturbed Forests vs. Logging Sites: (
A
,
E
) Shannon (H), (
B
,
E
) Dominance (D), (
C
,
G
)
Density (plants/km2) based on forest types, (D,H) Density (plants/km2) based on growth habit.
Forests 2023,14, 544 15 of 30
Table 4.
Comparison based on diversity, density, and distribution of life forms of orchid species in
undisturbed and disturbed forests in Peninsular Malaysia.
Parameters Undisturbed Forests Disturbed Forests
PG7 BR TRK BK KB LW TK BP BS GP BK(M) TM LHS KH LBK LB LG LP
Taxa_S 27 26 5 17 16 4 9 14 25 19 7 4 12 6 5 11 83 61
Occurrence (ni) 34 39 8 22 17 5 17 21 58 21 16 7 12 19 5 15 98 70
Occurrence (ni)
(Epiphytic orchids) 31 28 7 21 17 4 17 6 49 19 17 10 13 11 5 16 70 57
Occurrence (ni)
(Terrestrial orchids) 481221016200080000
Occurrence (ni)
(Lithophytic orchids) 0100000154000000000
Occurrence (ni)
(Mycoheterotroph
orchids)
100000000000000000
Dominance_D 0.047 0.049 0.250 0.066 0.066 0.280 0.142 0.098 0.092 0.057 0.258 0.306 0.083 0.191 0.200 0.102 0.015 0.018
Simpson_1-D 0.953 0.951 0.750 0.934 0.934 0.720 0.858 0.903 0.908 0.943 0.742 0.694 0.917 0.809 0.800 0.898 0.985 0.982
Shannon_H 3.200 3.139 1.494 2.776 2.752 1.332 2.069 2.491 2.801 2.912 1.629 1.277 2.485 1.709 1.609 2.338 4.342 4.063
Evenness_eˆH/S 0.909 0.888 0.891 0.944 0.979 0.947 0.879 0.863 0.658 0.969 0.728 0.897 1.000 0.920 1.000 0.942 0.926 0.953
Area (km2)0.900 14.400 16.800 5.618 5.500 3.600 12 9 11.760 34 12.600 0.250 13 0.800 33.300 1200 19.500 53
Density (plants/km2)37.800 2.700 0.500 3.900 3.100 1.400 1.400 2.300 4.900 0.600 1.300 28 0.900 23.800 0.200 0 5 1.300
Density of epiphytic
orchids (plants/km2)34.400 1.900 0.400 3.700 3.100 1.100 1.400 0.700 4.200 0.600 1.300 40 1 13.800 0.200 0 3.6 1.1
Density of terrestrial
orchids (plants/km2)4.400 0.600 0.100 0.400 0.400 0.300 0 0.100 0.500 0.100 0 0 0 10 0 0 0 0
Density of lithophytic
orchids (plants/km2)0 0.100 0 0 0 0 0 1.700 0.300 0 0 0 0 0 0 0 0 0
Density of
mycoheterotroph orchids
(plants/km2)
1.100000000000000000
3.3. Microclimate Conditions
Overall data of the microclimate conditions (Supplementary Table S3) show that
maximum temperature for the disturbed forests were significantly higher (p< 0.001,
df = 58) than the undisturbed ones. Simultaneously, for the relative air humidity, the
disturbed forests had significantly lower values (p< 0.001, df = 58) if compared to the
undisturbed forests (Table 5, Figure S2). Additionally, a comparison on microclimate con-
ditions in between two adjacent forests, an undisturbed forest (BK) and a forest margin
(BK(M)) located next to a logging site (LBK), was conducted to see whether there were
any significant differences in microclimate between the two forests contrasting in level of
disturbance. The result shows there were significant differences in maximum temperature
(p< 0.001, df = 38), minimum and maximum relative air humidity (p< 0.001, df = 38), and
minimum (p= 0.010, df = 38) and maximum (p< 0.001, df = 38) light intensity in between
the two adjacent forests (Table 5, Figure S3). It indicates that the disturbance caused by
adjacent logging projected extreme microclimate conditions on the environment. Also, the
result shows significant differences in temperature (p< 0.001, df = 38), relative air humidity
(p< 0.001, df = 38), and light intensity (p< 0.001, df = 38) in between the undisturbed forests
and extremely disturbed forests (logging areas) (Table 5, Figure S4). The canopy disruptions
caused by the selective-cut and clear-cut logging produced openings in the canopy, which
had significantly affected the temperature and the light intensity. The temperature could
reach up to 40
◦
C in the afternoon and consequently reduced the humidity to 51%. It
indicates extreme dryness in the logging areas, limiting growth and distribution of orchids.
Forests 2023,14, 544 16 of 30
Table 5.
Summary of topographic and ecological predictors (mean
±
SE) of undisturbed and
disturbed forests in Peninsular Malaysia.
Site Temperature (◦C) Relative Air Humidity
(1%–100%)
Light Intensity, PPFD
(µmol m−2s−1)Elevation (m)
Undisturbed vs. disturbed forests (p< 0.005, n= 30, df = 58)
Undisturbed forests Min: 27.8 ±0.3 aMin: 77.1 ±1.2 aMin: 23.8 ±3.3 a
100–600
Max: 31.2 ±0.2 aMax: 89.6 ±0.9 aMax: 171.7 ±18.8 a
Disturbed forests Min: 27.2 ±0.3 aMin: 64.5 ±2.3 bMin: 107.1 ±16 b
40–800
Max: 35.7 ±0.8 bMax: 88.7 ±1aMax: 456.5 ±29.6 b
Adjacent undisturbed vs. disturbed forests (p< 0.005, n = 20, df = 38)
Undisturbed forests Min: 30.6 ±0.1 aMin: 86.5 ±0.6 aMin: 21.5 ±2.2 a
100–200
Max: 30.7 ±0.1 aMax: 88.2 ±0.6 aMax: 125.8 ±17 a
Disturbed forests Min: 32.2 ±0.6 aMin: 44.8 ±1.2 bMin: 51.9 ±7.7 b
200–500
Max: 40.6 ±0.3 bMax: 75.8 ±1.5 bMax: 406.8 ±44.3 b
Undisturbed and extremely disturbed (logging sites) forests (p< 0.005, n = 20, df = 38)
Undisturbed forests Min: 28.1 ±0.3 aMin: 80.7 ±1aMin: 33.4 ±5a
100–600
Max: 31.3 ±0.3 aMax: 89 ±0.5 aMax: 190.9 ±24.7 a
Logging sites Min: 32.9 ±0.7 bMin: 51.8 ±2.4 bMin: 132.1 ±16.6 b
200–500
Max: 40.2 ±0.4 bMax: 79.5 ±1.7 bMax: 410.5 ±34.5 b
Notes: Different superscripts indicate significant differences (p< 0.05) in Mann–Whitney test.
Figure 8shows the hourly means for temperature and relative air humidity differed
between habitats of different disturbance regimes. Hourly means differed between dis-
turbance regimes during daytime from 10 am to 4 pm. The corresponding hourly tem-
perature for the logging sites were significantly higher at the consecutive period of day
(morning, H(1) = 26.38, p= 0.002, df = 2; afternoon, H(1) = 26.75, p< 0.001, df = 2; late
afternoon, H(1) = 20.33, p< 0.001, df = 2) if compared to the undisturbed forests (morning,
H(1) = 14.00, p= 0.002, df = 2; afternoon, H(1) = 13.10, p< 0.001, df = 2; late afternoon,
H(1) = 5.85, p< 0.001, df = 2). Similarly, the result shows significantly higher temperature
for the logging site (morning, H(1) = 12.63, p= 0.005, df = 2; afternoon, H(1) = 22.72,
p< 0.001, df = 2; late afternoon, H(1) = 19.35, p< 0.001, df = 2) than the non-logging site
disturbed forests (morning, H(1) = 5.78, p= 0.005, df = 2; afternoon, H(1) = 11.62, p< 0.001,
df = 2; late afternoon, H(1) = 5.33, p< 0.001, df = 2). The result also showed significant
difference for relative air humidity in between the logging sites (morning, H(1) = 6.56,
p= 0.011, df = 2; afternoon, H(1) = 5.00, p< 0.001, df = 2; late afternoon, H(1) = 5.44,
p< 0.001, df = 2) and undisturbed forests (morning, H(1) = 13.10, p= 0.011, df = 2; afternoon,
H(1) = 14.50, p< 0.001, df = 2; late afternoon, H(1) = 14.10, p< 0.001, df = 2). The result
was also with significant difference for relative air humidity in between the logging sites
(morning, H(1) = 7.06, p= 0.050, df = 2; afternoon, H(1) = 5.44, p< 0.001, df = 2; late
afternoon, H(1) = 5.33, p< 0.001, df = 2) and non-logging disturbed forests (morning,
H(1) = 11.94, p= 0.050, df = 2; afternoon, H(1) = 13.56, p< 0.001, df = 2; late afternoon,
H(1) = 13.67, p< 0.001, df = 2), except for morning time.
Forests 2023,14, 544 17 of 30
Forests 2023, 14, x FOR PEER REVIEW 18 of 32
Figure 8. Daily course of temperature and relative air humidity in undisturbed and disturbed forests
for different disturbance regimes in Peninsular Malaysia (2016–2020). Note: Each point represents
means of period of day, morning (10 am–12 pm), afternoon (12 pm–2 pm), and late afternoon (2 pm–
4 pm).
3.4. Factors Influencing Abundance of Epiphytic Orchids: Host Plants and Bark Moisture
Effects of microclimate change due to the anthropogenic and naturogenic activities
are now often examined based on the forest canopy species, especially epiphytic orchid
species that rely on the host plants. In this paper, the general host characteristics and bark
moisture that may influence the distribution and abundance of epiphytic orchids are
mainly explored. In total, 102 host individuals harbouring wild orchids were inspected in
all the studied line-transect plots. There are five most significant types that accumulate
abundant epiphytic orchids, Type 4, 6, 13, 18, and 19 (Table 6), with wide and interlocking
crown roofing the understorey part from sun exposure and dryness. Additionally, the
study highlights the importance of the host plants that possess bark texture Type 9 and 12
(Table 6), smooth to fissured textures. The result demonstrated that epiphytic orchid di-
versity and abundance is favoured by rough-barked host. Only 15 families of host species
were present in the study localities. Total host species was undetermined because a num-
ber of host individual or the fallen trees recorded in the logging sites were only identified
to family level due to the missing of important morphological characters, such as leaves
and flowers. In the undisturbed and non-logging disturbed forests, epiphytic orchids
were prominently abundant growing on Saraca, Sygyzium, and Neonauclea tree species that
were growing along the riparian areas. Many clumps of epiphytic orchids were also rec-
orded growing on large snags (dead standing trees) that were covered by moist substrate
(mosses and lichens) in riparian forests. On average, an individual host tree could have
harboured more than ten epiphytic orchid species. Species abundance of epiphytic orchids
correlated with the bark moisture and texture. A box plot analysis shows that host plants
having bark moisture 60%–80% generally harboured more epiphytic orchids than hosts
with 0%–20%, 20%–40%, 40%–60%, and 80%–100% bark moisture (Figure 9). Fifteen dis-
tinct bark textures were recorded and described based on the surface morphology (Table
7 and Figure 10). Using cluster analysis (Figure 11), the difference of bark textures of host
plants, those of undisturbed forest or non-logging disturbed forest or logging site were
distinguished. Bark textures are generally grouped well based on the vegetation type and
disturbance level. This suggests bark textures could be associated with the microclimate
conditions of the forest. Furthermore, this explains that the textures that potentially influ-
enced the abundance of epiphytic orchids were likely to be modulated by climate.
Figure 8.
Daily course of temperature and relative air humidity in undisturbed and disturbed forests
for different disturbance regimes in Peninsular Malaysia (2016–2020). Note: Each point represents
means of period of day, morning (10 am–12 pm), afternoon (12 pm–2 pm), and late afternoon (2 pm–4 pm).
3.4. Factors Influencing Abundance of Epiphytic Orchids: Host Plants and Bark Moisture
Effects of microclimate change due to the anthropogenic and naturogenic activities
are now often examined based on the forest canopy species, especially epiphytic orchid
species that rely on the host plants. In this paper, the general host characteristics and
bark moisture that may influence the distribution and abundance of epiphytic orchids are
mainly explored. In total, 102 host individuals harbouring wild orchids were inspected in
all the studied line-transect plots. There are five most significant types that accumulate
abundant epiphytic orchids, Type 4, 6, 13, 18, and 19 (Table 6), with wide and interlocking
crown roofing the understorey part from sun exposure and dryness. Additionally, the
study highlights the importance of the host plants that possess bark texture Type 9 and
12 (Table 6), smooth to fissured textures. The result demonstrated that epiphytic orchid
diversity and abundance is favoured by rough-barked host. Only 15 families of host species
were present in the study localities. Total host species was undetermined because a number
of host individual or the fallen trees recorded in the logging sites were only identified to
family level due to the missing of important morphological characters, such as leaves and
flowers. In the undisturbed and non-logging disturbed forests, epiphytic orchids were
prominently abundant growing on Saraca,Sygyzium, and Neonauclea tree species that were
growing along the riparian areas. Many clumps of epiphytic orchids were also recorded
growing on large snags (dead standing trees) that were covered by moist substrate (mosses
and lichens) in riparian forests. On average, an individual host tree could have harboured
more than ten epiphytic orchid species. Species abundance of epiphytic orchids correlated
with the bark moisture and texture. A box plot analysis shows that host plants having bark
moisture 60%–80% generally harboured more epiphytic orchids than hosts with 0%–20%,
20%–40%, 40%–60%, and 80%–100% bark moisture (Figure 9). Fifteen distinct bark textures
were recorded and described based on the surface morphology (Table 7and Figure 10).
Using cluster analysis (Figure 11), the difference of bark textures of host plants, those
of undisturbed forest or non-logging disturbed forest or logging site were distinguished.
Bark textures are generally grouped well based on the vegetation type and disturbance
level. This suggests bark textures could be associated with the microclimate conditions
of the forest. Furthermore, this explains that the textures that potentially influenced the
abundance of epiphytic orchids were likely to be modulated by climate.
Forests 2023,14, 544 18 of 30
Table 6.
Types of host plants harbouring orchids in the undisturbed and disturbed forests in
Peninsular Malaysia.
Type Habit Crown
Form Crown Class Substrate
Cover
Height
(m) Species
1 Tree Round Overtopped Basal ca. 3–6
Greenea corymbosa (Jack) Voigt (Rubiaceae), Streblus ilicifolius
(S.Vidal) Corner (Moraceae)
2 Tree Broad Codominant Trunk ca. 20–30 Castanopsis sp. (Fagaceae), Artocarpus elasticus Reinw.
ex Blume (Moraceae)
3 Tree Spreading Dominant Trunk ca. 40 Campnosperma auriculatum (Blume) Hook.f.
(Anacardiaceae), Castanopsis sp. (Fagaceae)
4 Tree Spreading Intermediate
Basal,
trunk,
branches
ca. 5
Ficus fistulosa Reinw. ex Blume (Moraceae),
Polyalthia cauliflora Hook.f. & Thomson (Annonaceae),
Goniothalamus sp. (Annonaceae), Tristaniopsis whiteana
(Griff.) Peter G.Wilson & J.T.Waterh. (Myrtaceae),
Syzygium cf. salictoides (Ridl.) I.M.Turner (Myrtaceae)
5 Tree Columnar Intermediate Basal,
trunk ca. 15–20 Syzygium sp. (Myrtaceae), Castanopsis sp. (Fagaceae)
6 Tree Spreading Codominant Basal ca. 20 Dipterocarpus elongatus Korth. (Dipterocarpaceae),
Dipterocarpus oblongifolius Blume (Dipterocarpaceae)
7 Tree Vase Intermediate
Basal,
trunk,
branches
ca. 15 Playmitra macrocarpa Boetl. (Annonaceae)
8 Tree Weeping Overtopped
Basal,
trunk,
branches
ca. 15 Hydnocarpus illicifolius King (Achariaceae)
9 Tree Globular Dominant
Basal,
trunk,
branches
ca. 7–10
Calophyllum wallichianum Planch. & Triana (Calophyllaceae),
Leptospermum polygalifolium Salisb. (Myrtaceae)
10 Tree Globular Intermediate Basal ca. 10 Macaranga gigantea (Rchb.f. & Zoll.)
Müll.Arg. (Euphorbiaceae)
11 Tree Round Intermediate Branches ca. 20 Goniothalamus sp. (Annonaceae)
12 Tree Oval Dominant Basal ca. 40 Ochanostachys amentacea Mast.(Olacaceae)
13 Tree Broad Intermediate Basal,
branches ca. 10
Neonauclea pallida subsp. malaccensis (Gand.) Ridsdale
(Rubiaceae), Saraca declinata Miq. (Fabaceae), Eurya cf.
acuminata DC (Pentaphylacaceae)
14
Liana
Leaflet
wholly
arranged
Intermediate Stem ca. 20 Spatholobus ferrugineus (Zoll. & Moritzi) Benth. (Fabaceae)
15
Liana
Spreading Overtopped Stem ca. 6 Apocynaceae
16
Fallen
tree Spreading Emergent Branches,
trunk ca. 50 Dipterocarpaceae
17
Fallen
tree Round Codominant Branches ca. 20–30 Ficus sp. (Moraceae), Artocarpus elasticus (Moraceae)
18
Fallen
tree Spreading Dominant Branches,
trunk ca. 40 Shorea sp. (Balau) (Dipterocarpaceae), Hopea sp.
(Dipterocarpaceae), Pentace eximia (Malvaceae)
19
Fallen
tree Broad Codominant Trunk ca. 30 Cynometra malaccensis Meeuwen (Fabaceae),
Parkia sp. (Fabaceae)
Forests 2023,14, 544 19 of 30
Forests 2023, 14, x FOR PEER REVIEW 21 of 32
Figure 9. Relation between total epiphyte species density (species per tree) and host plants’ bark moisture (bubble plot; N = 150). According to their colour, cold
to hot colour, bubbles reflect multiple values.
Figure 9.
Relation between total epiphyte species density (species per tree) and host plants’ bark moisture (bubble plot; N = 150). According to their colour, cold to
hot colour, bubbles reflect multiple values.
Forests 2023,14, 544 20 of 30
Forests 2023, 14, x FOR PEER REVIEW 22 of 32
Figure 10. Types of bark texture of host plants in undisturbed and disturbed forests in Peninsular
Malaysia: (A) Type 1, (B) Type 2, (C) Type 3, (D) Type 4, (E) Type 5, (F) Type 6, (G) Type 7, (H) Type
8, (I) Type 9, (J) Type 10, (K) Type 10, (L) Type 11, (M) Type 12, (N) Type 13, (O) Type 14, (P) Type
15.
Figure 10.
Types of bark texture of host plants in undisturbed and disturbed forests in Peninsular
Malaysia: (
A
) Type 1, (
B
) Type 2, (
C
) Type 3, (
D
) Type 4, (
E
) Type 5, (
F
) Type 6, (
G
) Type 7, (
H
) Type 8,
(I) Type 9, (J) Type 10, (K) Type 10, (L) Type 11, (M) Type 12, (N) Type 13, (O) Type 14, (P) Type 15.
Forests 2023,14, 544 21 of 30
Table 7.
Types of bark texture of the host plants harbouring epiphytic orchids in the undisturbed and
disturbed forests in Peninsular Malaysia.
Types Bark Textures Sites Forest Types
1 Fissured 1, 2, 3, 5, 8, 9, 10, 11, 12, 13, 14, 17, 18 Undisturbed, disturbed
2 Fissured, lenticeled 1 Disturbed
3 Fissured, peeled off 5, 6 Disturbed
4 Fissured, peeled off, soft 8 Undisturbed
5 Fissured, shedding 3 Disturbed
6 Peeled off, dry 1, 5, 13, 15, 16 Disturbed
7 Ridged, warty 5 Disturbed
8 Smooth 2, 4, 14 Undisturbed, disturbed
9 Smooth to fissured, moist 2, 3, 7, 17, 18 Disturbed
10 Smooth to fissured, warty 9, 12 Undisturbed
11 Smooth, lenticeled 2, 4, 8 Undisturbed
12 Soft, small fissured, warty 3, 5 Disturbed
13 Strongly fissured 9, 10, 11 Undisturbed
14 Strongly fissured, shedding 6, 13, 15, 16 Disturbed
15 Strongly fissured, soft, dry 1, 13 Disturbed
Forests 2023, 14, x FOR PEER REVIEW 23 of 32
Figure 11. Similarity of host plants (bark textures) amongst the studied vegetation alliances using
Bray–Curtis Similarity Coefficient.
4. Discussion
4.1. Ecological Implications of Current Findings on Orchid Diversity
Species richness, diversity, and abundance of the wild orchids differed in the differ-
ent localities. It is known that orchid species richness differs across a land-use intensity
gradient [3,63]. Generally, increment in the number of orchid individuals at each locality
is corresponding to the classical species-area relationship where larger areas tend to con-
tain larger numbers of species [64,65]. Moreover, in certain occasions, in extremely dis-
turbed and dry forests, orchids do not decrease in species diversity and density with in-
creasing disturbance, suggesting that orchids may indeed be comparatively disturbance-
resilient [41,50]. In the current study, high abundance of orchids in disturbed forests was
majorly comprised of epiphytic orchids whose occurrence was highly influenced by num-
ber of fallen trees, accessibility, and ease of human access for collections in the logged
forest areas [50,66,67]. The problem of greater access to fallen trees in logged areas seems
to be a methodological problem that appears to invalidate the ‘higher diversity’ found in
the disturbed forests. Despite the vast area, the number and distribution of wild orchids
in the deforested area were influenced by the number of host plants for epiphytic orchids
and the microclimate conditions. Additionally, in some logging sites, especially the ones
that have been neglected for many years, only one or two fallen trees are found over such
a large area. Moreover, since most of the ground vegetation has been cleared for skid
trails, ground orchids are rarely seen in the clearing zone and therefore have a low density.
As in this study, species density decreased with increased disturbance level. The micro-
climate of disturbed forests was markedly altered from that of undisturbed forests, simi-
larly, reported in [68] (also cited in [41]). Temperature, air humidity, and light intensity in
the undisturbed forests and the extremely disturbed forests (logging sites) were signifi-
cantly different proving that opening up of the forest canopy resulted in significant mi-
croclimate changes, which in turn affected the wild orchids [50,69].
Impoverishment of the logging sites was eminent, greatly contrasting with the non-
logging disturbed forests and the undisturbed forests. Logging sites were reported with
extremely high temperature and low relative air humidity and host plants’ bark moisture.
Figure 11.
Similarity of host plants (bark textures) amongst the studied vegetation alliances using
Bray–Curtis Similarity Coefficient.
4. Discussion
4.1. Ecological Implications of Current Findings on Orchid Diversity
Species richness, diversity, and abundance of the wild orchids differed in the different
localities. It is known that orchid species richness differs across a land-use intensity
gradient [
3
,
63
]. Generally, increment in the number of orchid individuals at each locality is
corresponding to the classical species-area relationship where larger areas tend to contain
larger numbers of species [
64
,
65
]. Moreover, in certain occasions, in extremely disturbed
and dry forests, orchids do not decrease in species diversity and density with increasing
disturbance, suggesting that orchids may indeed be comparatively disturbance-resilient [
41
,
50
].
In the current study, high abundance of orchids in disturbed forests was majorly comprised
of epiphytic orchids whose occurrence was highly influenced by number of fallen trees,
accessibility, and ease of human access for collections in the logged forest areas [
50
,
66
,
67
].
Forests 2023,14, 544 22 of 30
The problem of greater access to fallen trees in logged areas seems to be a methodological
problem that appears to invalidate the ‘higher diversity’ found in the disturbed forests.
Despite the vast area, the number and distribution of wild orchids in the deforested area
were influenced by the number of host plants for epiphytic orchids and the microclimate
conditions. Additionally, in some logging sites, especially the ones that have been neglected
for many years, only one or two fallen trees are found over such a large area. Moreover,
since most of the ground vegetation has been cleared for skid trails, ground orchids are
rarely seen in the clearing zone and therefore have a low density. As in this study, species
density decreased with increased disturbance level. The microclimate of disturbed forests
was markedly altered from that of undisturbed forests, similarly, reported in [
68
] (also cited
in [
41
]). Temperature, air humidity, and light intensity in the undisturbed forests and the
extremely disturbed forests (logging sites) were significantly different proving that opening
up of the forest canopy resulted in significant microclimate changes, which in turn affected
the wild orchids [50,69].
Impoverishment of the logging sites was eminent, greatly contrasting with the non-
logging disturbed forests and the undisturbed forests. Logging sites were reported with
extremely high temperature and low relative air humidity and host plants’ bark moisture.
The abundance of epiphytes usually decreases in sites with lower precipitation [
14
,
41
,
70
].
The smaller number of orchids’ taxa and terrestrial orchids in the non-logging disturbed
forests (excluding logging sites) than the undisturbed forests indicates that the study area
with extreme microclimate condition is a marginal habitat for wild orchids and seems to
confirm the importance of temperature and humidity [
41
,
70
–
72
], light intensity, and the
ground conditions [
50
] as a driver of wild orchids’ species richness. Dry microclimate in the
secondary forest, caused by the more open canopy of the secondary forest and the stronger
radiation drove impoverishment [
69
]. The authors of [
69
] report that high temperature, low
humidity, and lack of a dense bryophyte cover affected the bark moisture, and apparently
had cascading effects on the diversity of epiphytic orchids in these forests. Epiphyte species
should show stronger host specificity in habitats where climatic conditions are subopti-
mal for their performance because the modulating effect of tree traits is stronger under
such conditions [73].
Presumably, drier sites featured only the dominant, robust, and drought-tolerant
species [
41
,
50
,
74
] such as of Bulbophyllum and Dendrobium species. This explains the
dominance of these genera in both undisturbed and disturbed forests. The drought-tolerant
orchids tolerate drought instead of avoiding it [
75
]. They may be limited in their ecological
distribution by their inability to maintain a positive carbon balance during repeated cycles
of wetting and drying [
76
]. The lack of broad, thick leaves (thick cuticles) and trichomes
(which protect against radiation and reduce water loss by protecting against wind and heat)
are associated with their drought avoidance phenology. The plants produce roots at the
beginning of the wet season and shed them as the dry season approaches [
77
,
78
]. Some
plants possess storage organs such as pseudobulbs or pseudobulbous stem and rhizome to
store water [79].
The results also show that sites with little anthropogenic influence have higher species
densities than those with high anthropogenic influence. Primary or undisturbed forest
areas have more natural vegetation and a protected microclimate, are less polluted, less
disturbed by humans, including collecting pressures, and have more abundant host tree
species with suitable characteristics [
80
,
81
]. The contradicting results of orchid diversity
between undisturbed and disturbed forests from this study may also have been influenced
by the unseen (unmeasured) drivers such as microbiomes [
82
,
83
] and soil properties [
18
,
84
].
Based on the genera abundance, the orchid flora of disturbed forests was not dis-
tinct from the canopy-protected and undisturbed forests, but still, the species richness
and density were reduced. Forests that were either regenerating after clear-cutting or
surrounded by plantations generally show considerably reduced species number com-
pared to the matured and primary ones. Similar observations have been made in several
studies, in which secondary moist forests regenerating after clear-cutting generally have
Forests 2023,14, 544 23 of 30
significantly reduced epiphyte species abundance compared to mature forests (vascular
plants [
14
,
85
,
86
], bryophytes [
2
,
87
–
89
], epiphyte [
41
]). Three sites we studied, BS, GP, KH,
which were regenerated and secondary forests, showed a similar pattern. The authors
of [
63
,
81
] reported that species richness and abundance of epiphytic orchids gradually
declined significantly with increasing habitat modification from unmanaged via remnant
forest patches, agricultural land to very strong land-use intensity. This could indicate the
regeneration of wild orchid communities in the disturbed forests following disturbance
is very slow. However, this depends on how the forests are managed after cutting. For
instance, one of our studied sites, PG7, is a regenerated forest that has now gazetted as a
High Conservation Value Forest (HCVF) area, showing a considerably high diversity and
density of orchids compared to the other studied undisturbed and disturbed sites. Different
management practices influenced the orchid species and their composition and density.
4.2. Applications of the Current Findings for Informing Conservation and Restoration of Orchids
This investigation also gives prominence to the importance of the adjacent forests or
the remaining natural forests for restoration as the forests constitute biological sources
that enable succession from species-poor systems to highly diverse ones [
90
,
91
]. However,
the disturbance projected by the forest clearing affected the ecological conditions and
environment of the canopy-covered adjacent area. Forest clearing created forest margin or
forest edge, where a large area of the forests was opened and disturbed, and a number of
fallen trees cut during logging fell into the adjacent area. The adjacent area experienced
increased temperature and light intensity. Consequently, it had lower diversity and density,
too, if compared to the adjacent undisturbed forest. Similarly reported in studies [
41
,
92
,
93
],
edge habitats experienced increased solar radiation, air temperature, vapour pressure
deficit, and wind speed compared to the forest interior. Edge habitat fostered lower total
species richness than any other habitat, but was similar in composition to closed forest [
41
].
Projected warming threatens the persistence of a population unless these populations
possess a high micro-evolutionary potential that aid them in tolerating the environmental
changes [
94
]. This includes species or certain species from genera Bulbophyllum,Coelog-
yne,Cymbidium,Eria,Liparis,Oberonia,Paphiopedilum,Panisea,Pholidota, and Vanilla [
95
].
Based on the comparative study on ecological and forest structure in Bukit Batu Kapal,
Terengganu, a lowland to hill forest area, the undisturbed patches were home to abundant
orchids and mature tree species, whereas the adjacent logged and secondary forest was
composed of Bulbophyllum and Coelogyne species, and invasive and pioneer ‘non-orchid’
plant species. Bulbophyllum and Coelogyne species employed ‘drought escape’ adaptive
strategy by shedding leaves and roots in dry conditions and leaving dormant pseudobulbs
or stems to minimise transpiration.
The species composition of epiphytic orchids depends on several factors other than
temperature, humidity, and light intensity, such as host plant species and bark characteris-
tics as shown in [
15
,
63
]. Associations between epiphytic orchids and host characteristics
are among the main factors affecting the diversity and distribution of epiphytes [
3
]. Along
with altered microclimate, this has been attributed to reduced surface area for colonisation,
and the lack of late successional substrates on young, small, and fast-growing host trees.
In comparison with study [
87
], an analysis of tree-wide cortical moss diversity in primary
rainforests, specialists recovered more slowly than generalists. In addition to this, the
dominance of invasive tree and shrub species may restrict colonisation by epiphytic species.
Exotic species such Acacia, invasive species such as Bambusa, and native pioneer species
such as Macaranga have smooth bark texture, less favourable for the epiphyte colonisation,
including the mosses and lichens. In forest areas that have been subjected to disturbance,
Acacia,Macaranga, bamboos, and several species of early successional woody shrubs, such
as Melastomataceae are common. Other than that, Macaranga,Arenga, and Acacia have
a dominant presence in the secondary or converted forests. Epiphytic orchid specificity
towards certain or single species of host plants was rather weak, in fact, their preference
was more on the host characteristics regardless of species. According to studies [
54
,
96
], host
Forests 2023,14, 544 24 of 30
specificity exists when mutual trade-offs prevent a host-dependent species from adapting
equally well to the full range of conditions found in different host species. The same
species of orchids found on a variety of host plant species have similar characteristics in
undisturbed lowland forests. This is a ‘parallel host-specific scenario’, a term coined in [
54
],
in which all epiphyte species exhibit the same low or high performance on a particular tree
species. Moreover, previous studies of few and small populations of an epiphytic orchid
species increases the likelihood that limitation to a single tree species was simply due to
randomness in combination with host biases [
8
,
97
]. For instance, the epiphytic orchids
preferred host plant species, such as Tristaniopsis,Syzygium, and Eurya species growing
along the riparian area, but not in the dry habitat as in the heath vegetation and belukar
(bushes). In the riparian area, overtopped and intermediate trees were canopy-protected
and constantly moist with thick substrate covering the trunk and branches. Host plants
with smooth and dry bark texture, such as Leptospermum javanica,Calophyllum species, and
Tristaniopsis merguensis, were abundant in Padang 7 which is peaked with coastal heath
vegetation. Fewer epiphytes growing on these trees was not just because of the unsuitable
bark texture, but was also highly affected by the extreme hot and dry environment of the
heath vegetation. In fact, epiphytic orchids’ associations with many host bark textures
varied between localities indicating that some of the bark textures are locality-specific.
Moreover, certain host traits may change during ontogeny which, in turn, may lead to
differential ‘life stage biases’ [
54
]. These traits include leaf size and deciduousness [
98
] and
bark properties [99].
Although species specificity (host-orchid species specificity) was not proved here in
the current study, previous works have shown that epiphytes depend to some degree
on potential host species [
40
] and some orchid species are restricted to certain host tree
species [
63
,
100
,
101
]. In the protected forests, tall and large trees with Type 6 bark texture
(soft, small fissured, peeled off, warty) were common, but rare in the secondary and
intensively used forests. Tall and large trees in disturbed forests are often isolated. The
crowns of such isolated large trees are also isolated from each other, meaning that large
parts of the crowns are exposed to sun, favouring the orchids which use pseudobulbs as a
water reservoir [
4
]. This was also observed in the current study where only the Bulbophyllum
and Coelogyne were commonly dominating the isolated trees. Thus, large and old host tree
individuals and host plants with rough or fissured bark are essential for the protection of
epiphytic biodiversity.
However, not only the tree size but also the suitability of the stand to provide a
microhabitat for the epiphytes is important. Some epiphytic orchids are more common
in smaller trees and shrubs, similar to what was observed in [
81
,
102
]. Small trees with
spreading crown along the riparian areas would accumulate as much as the tall trees. High
abundance of species was not only observed on tall trees but also on smaller trees with
spreading crown along the riparian area because, similarly, they are exposed to light and
moisture. Host plants growing in the riparian area offer epiphytes specific microclimatic
conditions. Similarly, the authors of [
103
–
106
] observed that the most preferred tree species
occupied riparian areas, lower slopes, and valleys near swamps. These type of host plants
were observed with very low abundance in disturbed forests. As reported in study [
54
],
there are numerous instances of extreme basic host specificity that are not the typical ‘tree’
growth morphology. A notable example here is liana. Lianas such as Spatholobus species
also accumulate epiphytic orchids, but the orchids were not recorded as being abundant
as on trees.
Previous studies indicate that host species with rough bark are more frequently
colonised by epiphytes [
15
,
63
,
107
,
108
]. The reasons for more epiphytes occurring on
rough and fissured bark are that they retain moisture for longer and seedling recruitment
is better as seed lodges more easily in the crevices in bark than on smooth bark [
109
,
110
].
Host plants with either flaking, large peeling or smooth bark, and low substrate cover
are observed as poor hosts. The colonisation of bryophytes and lichens may influence
substrate suitability for vascular epiphytes [
8
,
71
,
111
,
112
]. However, quantitative evidence
Forests 2023,14, 544 25 of 30
is lacking to support this claim. Rarity of the epiphytic orchids in young secondary forests
was influenced by bark characteristics and tree architecture such as having little-branched
crowns with oblique instead of horizontal branches [
69
]. This type of crown architecture is
unfavourable for the establishment of epiphytes ([
113
]; cited in [
69
]). In secondary forests
in Peninsular Malaysia, trees with round and overtopped crowns with a height of 6 m to
10 m are common. The hosts’ bark texture significantly increased towards higher human
impact, showing significant differences between all intensities [
63
]. Previous studies in-
dicate that the diversity of epiphytes is positively correlated with host characteristics of
bark texture [
15
,
63
], type of host (deciduous or evergreen) [
114
], pH of bark, light, wind,
and tree diameter [
115
]. Unfortunately, in this study, the hosts’ ages and sizes (diameter at
breast height, tree height, and canopy volume) were not fully accounted for, as the age and
size of some individuals of the fallen trees and lianas are difficult to verify. Additionally,
the substrate cover was recorded only based on their general distribution on the hosts.
Terrestrial and mycoheterotroph orchids were found to be sensitive indicators of
microclimate and human disturbance in the disturbed forests. Since most of the terrestrial
and mycoheterotroph orchid species lack water storage organs and leaf cuticles, they are
highly dependent on the soil conditions, more sensitive to forest disturbance, and favoured
in forests with a dense tree canopy. Based on the findings, the terrestrial orchid species
were underrepresented in disturbed forests. Additionally, the predominance of lithophytic
orchids growing in rocky outcrop habitats, such as conglomerate hill and limestone hill
forests, is significant. Some orchid species in the rocky outcrops are capable of living in two
growth forms, growing as epiphytic and lithophytic. Many orchids, such as Bulbophyllum,
Coelogyne, and Dendrobium species, which are normally epiphytic, but also grow on the
ground or on limestone cliffs, similar to those reported in [
116
], are on the summit areas
with high light levels. Their adaptation increases the likelihood that the rocky outcrop is
lacking alternative hosts [
54
]. As reported in study [
117
], some shrubby or arborescent,
mat-forming monocots have been found as alternative hosts on rocky outcrops in Africa
and South America, including several Velloziaceae [
118
–
121
] and one Cyperaceae [
117
].
In this study, the alternative hosts could be bamboos (Poaceae) and Eugeissona or bertam
(Arecaceae) growing at the bottom of rocky outcrops, but wild orchids were not seen
growing there. Bases of limestone hills and ravines are usually shaded and have high
humidity, which often creates a microhabitat that is suitable for terrestrial orchids [
18
,
116
],
including Paphiopedilum niveum (Rchb.f.) Stein, Nervilia concolor (Blume) Schltr., and a
Crepidium species.
5. Conclusions
Conclusively, microclimatic changes, anthropogenic disturbance, and low abundance
of suitable host plants were key determinants of wild orchid composition and colonisation.
We suggest these criteria should be considered in forest restoration to reduce disturbance
impact on the associated orchid community. The study reported high diversity of epi-
phytic orchids and big differences in species richness, abundance, and species composition
of orchid communities in different localities across disturbance gradients in Peninsular
Malaysia. Wild orchids showed sensitivity to habitat quality and disturbance. Generally,
wild orchid species density was decreasing with increasing human impact, especially with
removal of host plants and extreme high temperature and light, and low humidity and
bark moisture. Occurrence of terrestrial and mycoheteretroph orchids are indicators of
microclimatic conditions and human disturbance. In non-logging forests, orchid abundance
changes along the disturbance gradient. Epiphytic assemblages exhibit correlation to the
type, bark texture, and bark moisture of the host plants. Conservation values of the suitable
host plants for epiphytic orchids’ diversification is worth further study. The significant
difference in microclimate of the adjacent undisturbed and disturbed forests appears to
underline the extreme effect of the disturbance which happened in the disturbed area
towards the intact adjacent area. Forest and ecological alterations caused a shift in patterns
of orchid species composition, where the robust and sun-loving species are dominating the
Forests 2023,14, 544 26 of 30
forests. However, the considerable effect of the disturbance towards the community in the
specific vegetation types requires further research because our current study was limited to
the pre-created trails only.
Supplementary Materials:
The following supporting information can be downloaded at:
https://www.mdpi.com/article/10.3390/f14030544/s1, Table S1: