Journal of

Plant Ecology A comparison of species

VOLUME 3, NUMBER 3,
PAGES 165–174 composition and stand structure
SEPTEMBER 2010

doi: 10.1093/jpe/rtq004 between planted and natural

mangrove forests in Shenzhen Bay,
Advanced Access published
on 3 March 2010

available online at
www.jpe.oxfordjournals.org
South China
Zhongkui Luo1,2, Osbert Jianxin Sun1,* and Hualin Xu3

Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010

1
MOE Key Laboratory for Silviculture and Conservation and Institute of Forestry and Climate Change Research, Beijing
Forestry University, Beijing 100083, China
2
State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing
100093, China
3
Neilingding-Futian National Nature Reserve of Guangdong, Shenzhen 518040, China
*Correspondence address. College of Forest Science, Beijing Forestry University, 35 Qinghua East Road, Beijing
100083, China. Tel: +86-10-62337095; E-mail: sunjianx@bjfu.edu.cn

Abstract

Aims
For assisting faster restoration of damaged or severely disturbed
coastal ecosystems, selected mangrove species have been planted
on previously mangrove-inhabited sites of the tropical and subtrop-
ical coasts of southern China. The objective of this study was to un-
derstand the stand dynamics of the planted mangroves and their
functional traits in comparison with natural mangrove forests under
similar site conditions.
Methods
Species composition, stand density, tree size distribution, and above-
ground production were investigated along three transects in a 50-
year-old planted mangrove stand and three transects in an adjacent
natural mangrove stand in Shenzhen Bay, South China. Measure-
ments were made on tree distribution by species, stand structure,
and aboveground biomass (AGB) distribution. Analyses were per-
formed on the spatial patterns of tree size distribution and species
association.
more variable in the planted stand than in the natural stand,
indicating higher spatiotemporal heterogeneity in the develop-
ment and succession of planted mangroves. Geostatistical analy-
ses show that both DBH and AGB were spatially auto-correlated
within a specific range in the direction perpendicular to coastline.
More than 60% of the variance in these attributes was due to
spatial autocorrelation. The Ripley’s K-function analysis shows
that the two dominant species, Kandelia obovata and Avicennia
marina, clumped in broader scales in the natural stand than in
the planted stand and displayed significant interspecific competi-
tion across the whole transect. It is suggested that interspecific
competition interacts with spatial autocorrelation as the underly-
ing mechanism shaping the mangrove structure. This study dem-
onstrates that at age 50, mangrove plantations can perform
similarly in stand structure, spatial arrangement of selected stand
characteristics and species associations to the natural mangrove
forests.

Important Findings Keywords: coastal ecosystem d forest restoration d mangroves

We found that the planted and natural mangrove stands did not d spatial autocorrelation d species association

differ in stand density, average diameter at breast height (DBH),

Received: 29 October 2009 Revised: 28 January 2010 Accepted: 8
species composition, and AGB. Spatial distribution of AGB and
February 2010
frequency at species level were also similar between the planted
and natural stands. However, the traits in stand structure were

Ó The Author 2010. Published by Oxford University Press on behalf of the Institute of Botany, Chinese Academy of Sciences and the Botanical Society of China.
All rights reserved. For permissions, please email: journals.permissions@oxfordjournals.org
166
INTRODUCTION
Mangroves are the only forests bridging the land and sea in the
tropical and subtropical coastal regions and play a vital role in
providing habitats for coastal animals and birds as well as in
serving as natural defense against disturbance (Alongi 2008;
Barbier et al. 2008). However, clearing for aquaculture and
other land uses has caused a sharp decline in mangrove forests
worldwide, such that over the past half century, more than
one-third of the mangrove forests have vanished, mainly
due to various anthropogenic activities (Alongi 2002; Valiela
et al. 2001). To save mangroves from extinction and to reverse
the trend of declining mangrove forests, planting has been
widely practiced to restore or rehabilitate this special type of
ecosystem (Ellison 2000; Field 1998; Walters 2003).
Despite being species poor, mangroves are one of the most
productive ecosystems (Kathiresan and Bingham 2001). Glob-
ally, the aboveground biomass (AGB) of mangrove forests
varies over a large range, from only 6.8 to as much as 436.4
Mg ha
1 (Saenger and Snedaker 1993), and aboveground
net primary productivity varies between 6.9 and 75.6 Mg
ha
1 year
1 (Sherman et al. 2003). The carbon accumulation
of the planted mangroves is comparable to that of the natural
mangroves (Field 1998; Kairo et al. 2008). For example, in a 7-
year-old Rhizophora mucronata plantation in Tritih, Indonesia,
the total AGB reached 94 Mg ha
1 (Sukardjo and Yamada
1992), whereas in a 12-year-old R. mucronata plantation at
Gazi Bay, Kenya, the standing biomass amounted to 107
Mg ha
1 (Kairo et al. 2008). Apart from the absolute value, that
how AGB is partitioned among species can affect ecosystem
functioning (Grime 1998) and indicate productivity, availabil-
ity and heterogeneity of resources and stand structure at com-
munity level (Edelkraut and Güsewell 2006; Luzuriaga et al.
2002; Parsons et al. 1994). Mangroves developed from different
planting schemes have been shown to display distinct alloca-
tion of biomass among species, leading to marked differences
in nutrient absorption and return rate (Liao et al. 1999).
Structural heterogeneity is an important ecosystem attri-
bute. Most of the mangrove species have specific zonation in
the intertidal area, which is one of the most important structural
features of mangrove forests and is the basis for the functioning
of mangrove ecosystems (Bunt 1996; Lugo 1980). In order to
understand the functional traits of mangroves, it is important
to examine their structural features. Analyses of the spatial
pattern of trees can provide valuable information on the un-
derlying processes, such as intra- and interspecific interaction,
and the effects of environmental heterogeneity. At local scales,
spatial pattern of natural mangrove stands closely links with
soil properties, disturbance history, and predation or attack
of propagules by understory animals (Ball 2002; Clarke and
Allaway 1993; Feller et al. 2002; Sousa et al. 2003). For planted
forests, a recent study on the upland forest ecosystems shows
that the spatial pattern of trees is largely the results of the sil-
vicultural method and the natural regeneration (Maestre et al.
Journal of Plant Ecology
2005). McKee and Faulkner (2000) found that site-specific as
well as regional or local differences in hydrology-related fac-
tors, such as salinity and soil waterlogging, could have a strong
influence on the functional outcomes of planted mangroves in
their early development. The underlying mechanisms control-
ling the development of mangroves can be greatly influenced
by anthropogenic disturbance; this influence has been shown
to last for >20 years (Ferwerda et al. 2007).
In recent years, large areas of mangrove forests have been
established with selected mangrove species on previously
mangrove-inhabited sites of the tropical and subtropical coasts
of southern China. The practice is aimed for facilitating faster
restoration of coastal ecosystem structure and functions. Nu-
merous studies have been conducted to assess the structure
and function of natural mangroves and relatively young stands
Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010
of planted mangroves (e.g. <10 years) in the coasts of southern
China (see reviews by Lin 1997 and Chen et al. 2009). Zan et al.
(2001) found that mangrove plantations had fast growth rates
and varied greatly in species composition and stand structure
during the first several years after planting. However, informa-
tion is lacking on how well the planted mangroves may per-
form compared with natural mangrove forests and whether
the planted mangroves can reach the same stand structural
complexity as the natural mangroves for a given time, espe-
cially after several decades of planting. To address these con-
cerns, we studied species composition, stand density, tree size
distribution, and aboveground production along three trans-
ects in a 50-year-old planted mangrove stand and three trans-
ects in an adjacent natural mangrove stand in Shenzhen Bay,
South China. Geostatistical tools were used to analyze the spa-
tial patterns of tree size in terms of AGB and diameter at breast
height (DBH); Ripley’s K-functions were used to examine
associations of mangrove species. We mainly focused on
two questions: (i) what is the spatial relationship among
and within mangrove species in both planted and natural
mangrove stands; and (ii) would the planted mangroves de-
velop similar structural complexity to the natural mangroves
after 50 years of planting?
MATERIALS AND METHODS
Study site and field survey
This study was carried out in the Futian National Nature Re-
serve (22°32# N, 114°00# E), Shenzhen Bay, South China. The
Reserve stretches ;11 km along the coast and covers an area of
>300 ha, of which ;110 ha are natural mangrove forests (Tam
et al. 1995). The site has a tide that is semidiurnal with a mean
range of 1.4 m and maximum ;2.8 m at high spring tide. Av-
erage salinity of seawater is <15&, with a pH 7.6. The annual
mean temperature for the period 1953–2005 averaged 22.1°C,
and the minimum and maximum temperatures were 0°C and
39.1°C; and the mean annual precipitation was ;1 726 mm, of
which 70.7% fall between May and September.
Field survey and biological measurements were made be-
tween October 2005 and June 2006. We selected two
Luo et al. | Species composition and stand structure of mangroves
neighboring mangrove stands representing the natural man-
groves and the planted mangroves, respectively. The natural
mangrove stand is dominated by Avicennia marina (Forsk.)
Vierh and Kandelia obovata Sheue (K. obovata was formerly
named as Kandelia candel in Eastern Asia, see details in Sheue
et al. 2003), with a small proportion of Aegiceras corniculatum
(L.) Blanco and Bruguiera gymnorrhiza (L.) Lamk. Age of the
natural stand at the time of this study was ;60 years. The
planted mangrove stand was originally a K. obovata monocul-
ture plantation. K. obovata trees were hand-planted in 1950s
and the stand was not subject to any management scheme
since establishment (Zhang 1997). Subsequently, more man-
grove species, including exotic species such as Sonneratia
apetala Buch. -Ham. and Sonneratia caseolaris (L.) Engl, were
introduced into the plantation (Zhang 1997). Currently there
are >80 ha of mangrove plantations in the Reserve. Both types
of stands distribute along the coastline without vertical strat-
ification, with understory largely open except presence of
very few seedlings and some Acanthus ilicifolius L. (small shrub)
individuals.
Three 10-m wide transects were sampled from the upland
to the sea for each of the two stand types. These transects
were systematically selected and spaced out to cover the
spectrum of each stand type. All the transects were perpen-
dicular to coastline and traversed the whole mangroves.
On average, the adjacent transects were separated by
>100 m. Within each transect, species and location,
DBH and average canopy height were recorded for all live
trees.
Calculation of stand structure and AGB distribution
We estimated AGB using previously-developed allometric
equations for each of the four species, specifically, allometric
equations of Tam et al. (1995) for K. obovata, A. marina, Ae.
corniculatum and allometric equation of Weng (1999) for B.
gymnorrhiza. Regression analysis was used to assess the distribu-
tion of AGB along a transect from the upland to the sea. Pielou
evenness index (E) and Shannon–Wiener index (H) were used
to describe the diversity traits of plant communities based
on species frequency (F) and AGB. E and H were calculated
as follows:

S .

E = 1
+ ðNi =NÞ2 1
1=S ; ð1Þ
i=1
and
S

H =
+ ððNi =NÞ 3 ln Ni =N ; ð2Þ
i=1
where, S is the number of species, N is the number of individ-
uals or AGB for all species and Ni the number of individuals or
AGB for species i. The relative proportion (%) of AGB was also
calculated for each species.
A geometric series model (GS; also known as niche preemp-
tion model) was used to fit the abundance distribution pattern
167
of species in the two mangrove stand types. The GS was
calculated as follows (Zhang 2004):
yi = kð1
kÞ1
i ; ð3Þ
where, yi is the ith abundant species and k is the niche preemp-
tion parameter, which is a constant for a given community.
This model assumes that the first species colonizing the com-
munity preempts the first k fraction of the total resource or
space and the second species takes the k fraction of the remain-
der and so on, until the entire niche space is fully occupied.
Assessments of spatial patterns of tree size (DBH
and AGB)
The spatial characteristics of DBH and AGB in each transect
were examined with semivariogram. The experimental semi-
Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010
variogram, g(h), is derived by calculating one-half the variance
of data values for every pair of data locations along a specified
direction. These values were plotted against the distances be-
tween the data pairs. This can be expressed as follows:
N
gðhÞ = ð1=2NÞ + ½Zðxi Þ
Zðxi + hÞ; ð4Þ
i=1
where, xi is a datum location, h a lag vector, Z(xi) the data value
at location xi, Z(xi + h) the datum value at location xi + h, and N
the number of data pairs spaced h units apart. Here, we calcu-
lated the specific directional semivariogram perpendicular to
coastline. All the calculations were performed using GS+ for
window version 5 (Gamma Design Software 2004). For each
transect, the lag interval was 1 m, and the active-lag interval
was half of the transect length. To ensure normality, data were
log-transformed before analyses.
The experimental semivariogram depicts the spatial depen-
dence of studied variables over the sampling space. By fitting
the experimental semivariograms to the geostatistical model,
we derived information on the following: (i) the ‘sill’ (C0 + C),
which is the variogram value at which the plotted points level
off; (ii) the ‘nugget’ (C0), which reflects the spatial variability at
distances smaller than the smallest sampling distance; (iii) the
‘structure variance’ (C), which is the difference between the sill
and nugget; and (iv) the ‘range’ (A), which defines the average
distance within which the samples remain spatially correlated.
The ratio of C:(C0 + C) represents the proportion of total sample
variance induced by spatial autocorrelation.
Computation of species associations
For the two dominant species, K. obovata and A. marina, we
used Ripley’s K-function (Diggle 1983; Ripley 1977), K(r),
and the related functions of second-order neighborhood anal-
ysis to examine the intra- and interspecific associations. For
univariate analysis (i.e. intraspecific association), K(r) is com-
monly presented as the linearized L-function (Besag 1977):
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
LðrÞ = KðrÞ=p
r; ð5Þ
which is easier to interpret than K(r). At scale r, target species
are completely randomly distributed if L(r) = 0, clumped if L(r)
> 0 and regularly distributed if L(r) < 0.
168
L12(r), the transformation of bivariate Ripley’s K-function
K12(r), was used to analyze the association between K. obovata
and A. marina. L12(r) was calculated as follows (Besag 1977):

 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
L12 r = K12 ðrÞ=p
r; ð6Þ
At scale r, two species are spatially independent if L12(r) = 0,
negatively associated if L12(r) <0 and positively associated if
L12(r) > 0. For both univariate and bivariate analyses, 99%
confidence envelopes were calculated under the null hypoth-
esis of complete spatial random by Monte Carlo simulations
with 1 000 iterations. All the calculations and simulations were
performed using ADE-4 package (Thioulouse et al. 1997;
http://pbil.univ-lyon1.fr/ADE-4-old/ADE-4.html).
RESULTS
Stand structure and AGB distribution
The two types of mangrove stands did not show significant differ-
ences in average stand density, DBH and AGB (Table 1). The Pie-
lou evenness and Shannon–Wiener indices, which were
calculated based on species abundance and AGB, were also not
significantly different between the two stand types (Table 1).
The distribution of DBH was skewer in the planted stand than
in the natural stand (Fig. 1). The GS model well describes the
abundance distribution of the four species in both types of
stands (Fig. 2). The niche preemption parameter k was similar
between the two stand types, which was 0.717 for the planted
stand and 0.716 for the natural stand. Although all these struc-
tural variables did not show significant differences between the
planted and natural stands, the values of coefficient of variance
for the above variables were consistently greater in the planted
stand than in the natural stand except DBH (Table 1; Fig. 2).
On average, the planted stand extended 34.3 m seaward
from the upland, while the natural stand extended 43.3 m sea-
ward from the upland (Table 1; Fig. 3). The total AGB signif-
icantly (P < 0.01) and exponentially declined from upland to
sea in both the planted and natural stands (Fig. 3). Kandelia
obovata and A. marina differed in spatial distributions of
AGB: ;84.1% of AGB for K. obovata in the planted stand
and 73% in the natural stand distributed in the section 0–
15 m from the stand edge in upland, whereas the AGB for
A. marina extended more into the sea (Fig. 3).
Spatial autocorrelation of DBH and AGB
The experimental semivariograms for DBH and AGB were
more variable in the planted stand than in the natural stand
Table 1: stand characteristics (mean 6 SD, n = 3) of planted mangroves (PM) and natural mangroves (NM) in Shenzhen Bay, South China

2
Forest type Transect length (m) Stand density (plants/100 m ) DBH (cm)
PM 34.30 6 1.04 38 6 16 8.75 6 0.42 9.17 6 3.39
NM 43.30 6 10.91 32 6 3 9.26 6 0.96 9.01 6 2.07
Journal of Plant Ecology
(Fig. 4). In general, the linear model fitted the data well
(Table 2). Spatial autocorrelation accounted for 74.3%
and 70.3% of the variance in DBH and AGB, respectively,
for the planted stand, whereas in the natural stand, 73.4%
and 62.5% of the variance in DBH and AGB, respectively,
correlated with the spatial separation of samples (Table 2).
Species association
Avicennia marina showed similar spatial association between
the planted and natural stands but also displayed apparently
different spatial association among transects for the same
stand type (Fig. 5). Along the three transects for the planted
stand, A. marina was generally randomly distributed with
exception of significant and positive association at scales
Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010
7.5–11.5 m in one of the transects (Fig. 5a) and 5–15.5 m
in another (Fig. 5c). In the three transects for the natural
stand, A. marina was significantly and positively associated
at broader scales compared with the planted stand, at
3–14 m (Fig. 5d) and 1–23 m (Fig. 5f) in two different trans-
ects, respectively.
K. obovata not only showed similar spatial association be-
tween the planted and natural stands but also among transects
for the same stand type (Fig. 6). In the three transects for the
planted stand, K. obovata was significantly and positively asso-
ciated at scales up to 11 m, 10.5 m and 15.5 m (Fig. 6a–c), re-
spectively, whereas in the three transects for the natural stand,
K. obovata was significantly and positively associated at scales
up to 16.5 m, 13 m and 22.5 m (Fig. 6d–f), respectively. Of
these scales, the positive association changed into random as-
sociation (Fig. 6a–c, e–f) and then into significant and negative
association at scales >12.5 m and >13 m for two of the trans-
ects in the planted stand (Fig. 6a–b) and >15 m for one of the
transects in the natural stand (Fig. 6e).
Generally, the association between A. marina and K. obo-
vata was significant and negative at all scales along the whole
transect (Fig. 7). In the three transects for the planted stand,
the negative association of the two species occurred at scales
up to 11.5 m, 17.5 m and 17 m, respectively (Fig. 7a–c),
whereas in the three transects for the natural stand, the neg-
ative association of the two species occurred at scales up to 15
m, 16.5 m and 27.5 m, respectively (Fig. 7e–f). The two spe-
cies showed a random association at scales between 11.5 m
and 14 m, positive association at scales >14 m for one of
the transects in the planted stand (Fig. 7a) and random asso-
ciation at scales >15 m for one of the transects in the natural
stand (Fig. 7d).
Pielou evenness index Shannon–Wiener index

2
AGB (kg m ) Frequency AGB Frequency AGB
0.55 6 0.28 0.52 6 0.26 0.31 6 0.15 0.26 6 0.07
0.62 6 0.17 0.48 6 0.15 0.35 6 0.05 0.25 6 0.05
Luo et al. | Species composition and stand structure of mangroves
Figure 1: relative frequency distribution of DBH in plantation and
natural mangroves in Shenzhen Bay, South China.
Figure 2: fitting abundance of four mangrove species to GS model in
plantation and natural mangroves, in Shenzhen Bay, South China.
Vertical bars are the standard deviations (n = 3).
DISCUSSION
In this study, we assessed the stand structure, AGB, and species
associations in mangrove forests in Shenzhen Bay, South
China, and compared these characteristics between planted
and natural stands. Our results indicate that the planted man-
groves can be comparable to the natural mangroves in stand
structure after 50 years of planting and that the species distri-
bution and tree size (DBH and AGB) are spatially auto-corre-
lated to some extent. K. obovata and A. marina dominate the
mangrove forests of the study site with their own separate zo-
169
Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010
Figure 3: AGB distribution of four mangrove species along transect
perpendicular to coastline in plantation and natural mangroves in
Shenzhen Bay, South China. Vertical bars are the standard deviations
(n = 3).
nation, which could be explained by significant intraspecific
clumping and interspecific competition.
Forest structure and function vary with development and
succession, and the planted and natural forests generally have
respective characteristicsin differentdevelopmentstages(Barlow
et al. 2007; Guariguata and Ostertag 2001; Lugo 1992). Proffitt
and Devlin (2005) found that an 11-year-old Rhizophora mangle
plantation was comparable with the natural mangrove stand in
species richness, vegetation cover, litterfall and light penetration
but had considerably smaller trees and higher stand density. In
this study, we also did not find significant differences in stand
structure, AGB, resource allocation and spatial tree distribution
between planted and natural mangroves at age ;50 (Fig. 2),
similar to the findings of Lugo (1992) comparing 50-year-old
planted and natural mangroves. However, the planted man-
grove stand was more variable in specific stand characteristics
than the natural mangrove stand in our study. This greater var-
iation in the planted mangrove stand reflects high spatiotempo-
ral heterogeneity in the development and succession of planted
mangroves and indirectly demonstrates the sensitivity of plan-
tation to fluctuation of environmental conditions, such as cli-
matic disturbance and pathogenic outbreak, during succession.
Different mangrove species have distinct zonations. This
unique characteristic of mangrove species may significantly in-
fluence the structure and function of monocultural mangrove
plantation with succession. In this study, the K. obovata planta-
tion was exclusively dominated by K. obovata along the stand
edge in upland after ;50 years of development, while A. marina
dominated the part of the transects extending into the sea. Fur-
thermore, the planted mangrove stand extended less into the
sea than the natural mangrove stand, albeit similar species com-
position and stand structure between the two types of stands.
170 Journal of Plant Ecology

Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010

Figure 4: experimental semivariograms for DBH and AGB in three planted mangrove transects (PM1, PM2 and PM3) and three natural mangrove
transects (NM1, NM2 and NM3) in Shenzhen Bay, South China.

Table 2: correlative parameters* for the spatial patterns of DBH and AGB in three plantation mangrove transects (PM1, PM2 and PM3) and
three natural mangrove transects (NM1, NM2 and NM3) in Shenzhen Bay, South China

Variable Forest type Model Nugget (C0 3 102) Sill ((C0 + C) 3 102) Range (A) C/(C0 + C) R2

DBH PM1 Linear 9.30 68.28 25.00 0.86 0.87

PM2 Linear 23.70 72.95 80.72 0.68 0.22
PM3 Exponential 28.30 91.04 118.08 0.69 0.27
NM1 Linear 5.20 25.21 38.40 0.79 0.76
NM2 Linear 3.26 10.56 52.24 0.69 0.61
NM3 Linear 1.48 5.24 64.73 0.72 0.71
t-test — 0.042 0.002 0.462 0.909 —
AGB PM1 Spherical 26.50 91.39 53.43 0.71 0.83
PM2 Exponential 2.75 9.56 60.81 0.71 0.71
PM3 Linear 73.50 235.15 724.50 0.69 0.003
NM1 Linear 37.70 109.92 147.40 0.66 0.14
NM2 Linear 36.70 105.10 480.50 0.65 0.03
NM3 Linear 26.10 59.95 74 650.00 0.57 0.01
t-test — 0.973 0.782 0.373 0.068 —

*See text for explanations on nugget, sill, range and C/(C0 + C).

Tree size distribution is an important attribute reflecting for-
est development and successional stage. Our analyses of the
distribution of DBH and AGB, which characterize the tree size,
indicate that in the direction perpendicular to coastline, the
two variables are apparently spatially structured and the aver-
age range of spatial correlation is larger than the whole tran-
sect length. Moreover, the experimental semivariogram of
these variables is linear. These results are consistent with stud-
ies in upland forests and indicate that neighborhood competi-
tion plays an important role in regulating spatial structure of
tree size (Biondi et al. 1994; Kuuluvainen et al. 1998; Nanos
et al. 2004). The planted mangrove stand displayed greater var-
iance of semivariogram, and greater positive skew in tree size
distribution than the natural mangrove stand, demonstrating
Luo et al. | Species composition and stand structure of mangroves 171

Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010

Figure 5: univariate analyses of distribution patterns (solid line) of Avicennia marina in three transects for the planted mangroves (a, b and c) and
three transects for the natural mangroves (d, e and f) using linearized L-function, L(r), in Shenzhen Bay, South China. The two dash lines represent
the corresponding 99% confidence envelopes of random distribution.

Figure 6: univariate analyses of distribution patterns (solid line) of Kandelia obovata in three transects for the planted mangroves (a, b and c) and
three transects for the natural mangroves (d, e and f) using linearized L-function, L(r), in Shenzhen Bay, South China. The two dash lines represent
the corresponding 99% confidence envelopes of random distribution.
172
Figure 7: bivariate analysis of interactions between Avicennia marina and Kandelia obovata in three transects for the planted mangroves (a, b and c)
and three transects for the natural mangroves (d, e and f) using the transformation of bivariate Ripley’s K-function, L12(r), in Shenzhen Bay, South
China. The two dash lines represent the corresponding 99% confidence envelopes of random distribution.
the greater neighborhood competition or self-thinning in the
planted mangroves. Compared with the natural mangrove
stand, the planted stand was characterized by having greater
number of smaller trees after 50 years of development, indicat-
ing the early stage of stand development. Liao et al. (2004)
found that the mangrove communities at our study site were
under a strong influence of self-thinning, suggesting the oc-
currence of a dynamic stand developmental process.
Similar to many upland tree species (Condit et al. 2000; Wolf
2005), the two dominant mangrove species in our study, K.
obovata and A. marina, are clumped to some extent. There
are apparent differences in clump intensity between the two
species but not between the two stand types. Negative associ-
ations between K. obovata and A. marina generally reflect the
interspecific competition (Rejmánek and Lepš 1996), which
explains the mutually exclusive and specific zonations of
the two species. Studies of the underlying mechanisms that
control this phenomenon of zonation have traditionally fo-
cused on the effects of light condition (Ball 2002; Clarke
and Allaway 1993), soil heterogeneity (Chen and Twilley
1999; Ellison et al. 2000; Feller et al. 2002; Matthijs et al.
1999; McKee 1993), crab predation on propagules (Clarke
and Kerrigan 2002; Smith 1987), or a combination of these fac-
tors, on early growth of mangrove species. Our study did not
directly or experimentally examine the causes of the observed
zonation pattern of mangrove species. We may postulate that it
is driven by the different resource availability or the soil het-
Journal of Plant Ecology
Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010
erogeneity from the upland to the sea. The comparable zona-
tion of mangrove species between planted and natural
mangroves indirectly supports the assumption that propagule
availability cannot account for the zonation of mangrove spe-
cies (McKee 1995; Sousa and Mitchell 1999; Sousa et al. 2007).
Because of the great economic and ecological value of man-
grove forests (Barbier et al. 2008; Han et al. 2000), plantations
have become one of the most important ways for the restora-
tion of this threatened forest type in the tropical and subtrop-
ical coastal regions. Whether this artificially created ecosystem
can serve the same ecological role (e.g. resistance to distur-
bance, carbon fixation and biodiversity conservation) as the
natural ecosystems largely depends on its structure and func-
tion. Although there are greater variations in the attributes of
planted mangrove stand than of the natural mangrove stand,
our results indicate that at age 50, the planted mangroves are
comparable in stand structure, species association, and AGB to
the natural mangroves. Combining with suitable management
approaches (Lewis 2005), reforestation could be a feasible and
effective way to not only reverse the declining trend of man-
groves but also restore the structure and function of mangrove
ecosystems.
FUNDING
Key Knowledge Innovation Program of the Chinese Academy
of Sciences (KSCX2-SW-132); ‘11th Five-Year’ Forestry
Luo et al. | Species composition and stand structure of mangroves
Scientific and Technological Support Program of the Ministry
of Science and Technology of China (2008BADB0B0302).
ACKNOWLEDGEMENTS
We thank Z.-Y. Zhou, L.-Z. Chen and X.-Q. Zeng for field assistance
and Futian Mangrove Natural Reserve for site access permission and
technical support.
Conflict of interest statement. None declared.
REFERENCES
Alongi DM (2002) Present state and future of the world’s mangrove
forests. Environ Conserv 29:331–49.
Alongi DM (2008) Mangrove forests: resilience, protection from tsu-
namis, and responses to global climate change. Estuar Coast Shelf
Sci 76:1–13.
Ball MC (2002) Interactive effects of salinity and irradiance on growth:
implications for mangrove forest structure along salinity gradients.
Trees 16:126–39.
Barbier EB, Koch EW, Silliman BR, et al. (2008) Coastal ecosystem-
based management with nonlinear ecological functions and values.
Science 319:321–3.
Barlow J, Gardner TA, Araujo IS, et al. (2007) Quantifying the biodi-
versity value of tropical primary, secondary, and plantation forests.
Proc Natl Acad Sci U S A 47:18555–60.
Besag J (1977) Contribution to the discussion of Dr. Ripley’s paper. J R
Stat Soc B 39:193–5.
Biondi F, Myers DE, Avery CC (1994) Geostatistically modeling stem size
and increment in an old-growth forest. Can J Forest Res 24:1353–68.
Bunt JS (1996) Mangrove zonation: an examination of data from sev-
enteen riverine estuaries in tropical Australia. Ann Bot 78:333–41.
Chen R, Twilley RR (1999) Patterns of mangrove forest structure and
soil nutrient dynamics along the Shark river estuary, Florida. Estu-
aries 22:955–70.
Chen L, Wang W, Zhang Y, et al. (2009) Recent progresses in mangrove
conservation, restoration and research in China. J Plant Ecol 2:45–54.
Clarke PJ, Allaway WG (1993) The regeneration niche of the grey
mangrove (Avicennia marina): effects of salinity, light and sediment
factors on establishment, growth and survival in the field. Oecologia
93:548–56.
Clarke PJ, Kerrigan RA (2002) The effects of seed predators on the re-
cruitment of mangroves. J Ecol 90:728–36.
Condit R, Ashton PS, Baker P, et al. (2000) Spatial patterns in the dis-
tribution of tropical tree species. Science 288:1414–8.
Diggle PJ (1983) Statistical Analysis of Spatial Point Patterns. New York:
Academic Press.
Edelkraut KA, Güsewell S (2006) Progressive effects of shading on ex-
perimental wetland communities over three years. Plant Ecol
183:315–27.
Ellison AM (2000) Mangrove restoration: do we know enough? Restor
Ecol 8:219–99.
Ellison AM, Mukherjee BB, Karim A (2000) Testing patterns of zona-
tion in mangroves: scale dependence and environmental correlates
in the Sundarbans of Bangladesh. J Ecol 88:813–24.
Feller IC, McKee KL, Whigham DF, et al. (2002) Nitrogen vs. phospho-
rus limitation across an ecotonal gradient in a mangrove forest. Bio-
geochemistry 62:145–75.
173
Ferwerda JG, Ketner P, Mcguinness KA (2007) Differences in regen-
eration between hurricane damaged and clear-cut mangrove stands
25 years after clearing. Hydrobiologia 591:35–45.
Field CD (1998) Rehabilitation of mangrove ecosystems: an overview.
Mar Pollut Bull 37:383–92.
Gamma Design Software GS+: Geostatistics for the Environmental Sciences.
Plainwell, MI: Gamma Design Software.
Grime JP (1998) Benefits of plant diversity to ecosystems: immediate,
filter and founder effects. J Ecol 86:902–10.
Guariguata MR, Ostertag R (2001) Neotropical secondary forest suc-
cession: changes in structural and functional characteristics. Forest
Ecol Manage 148:185–206.
Han W, Gao X, Lu C, et al. (2000) Ecological value assessment of man-
grove forest ecosystems in China. Ecol Sci 19:40–5(in Chinese with
English abstract).
Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010
Kairo JG, Lang’at JKS, Dahdouh-Guebas F, et al. (2008) Structural de-
velopment and productivity of replanted mangrove plantations in
Kenya. Forest Ecol Manage 225:2670–7.
Kathiresan K, Bingham BJ (2001) Biology of mangroves and man-
grove ecosystems. Adv Mar Biol 40:81–251.
Kuuluvainen T, Järvinen E, Hokkanen TJ, et al. (1998) Structural het-
erogeneity and spatial autocorrelation in natural mature Pinus syl-
vestris dominated forest. Ecography 21:159–74.
Lewis RR III (2005) Ecological engineering for successful management
and restoration of mangrove forests. Ecol Eng 24:403–18.
Liao WB, Lan CY, Zan QJ, et al. (2004) Growth dynamics of self-thin-
ning of the dominant populations in the mangrove community. Acta
Bot Sin 10:522–32.
Liao WB, Zheng DZ, Li YD (1999) Aboveground biomass and nutrient
accumulation and distribution in different type Sonneratia carseo-
laria–Kandelia candel mangrove plantations. Chinese J of Appl Ecol
10:11–5(in Chinese with English abstract).
Lin P (1997) Mangrove Ecosystems in China. Beijing: Science Press 271.
Lugo AE (1980) Mangrove ecosystems: successional or steady state?
Biotropica 12:65–72.
Lugo AE (1992) Comparison of tropical tree plantations with second-
ary forest of similar age. Ecol Monogr 62:1–41.
Luzuriaga AL, Escudero A, Loidi J (2002) Above-ground biomass dis-
tribution among species during early old-field succession. J Veg Sci
13:841–50.
Maestre FT, Escudero A, Martı́nez I, et al. (2005) Does spatial pattern
matter to ecosystem functioning? Insights from biological soil
crusts. Func Ecol 19:566–73.
Matthijs S, Tack J, van Speybroeck D, et al. (1999) Mangrove species
zonation and soil redox state, sulphide concentration and salinity in
Gazi Bay (Kenya), a preliminary study. Mangroves Salt Marshes
3:243–9.
McKee KL (1993) Soil physicochemical patterns and mangrove species
distribution—reciprocal effect? J Ecol 81:477–87.
McKee KL (1995) Mangrove species distribution and propagule preda-
tion in Belize: an exception to the dominance-predation hypothesis.
Biotropica 27:334–45.
McKee KL, Faulkner PL (2000) Restoration of biogeochemical func-
tion in mangrove forests. Restor Ecol 8:247–59.
Nanos N, González-Martı́nez SC, Bravo F (2004) Studying within-
stand structure and dynamics with geostatistical and molecular
marker tools. Forest Ecol Manage 189:223–40.
174
Parsons AN, Welker JM, Wookey PA, et al. (1994) Growth-responses of
4 sub-arctic dwarf shrubs to simulated environmental-change. J
Ecol 82:307–18.
Proffitt CE, Devlin DJ (2005) Long-term growth and succession in
restored and natural mangrove forests in southwestern Florida.
Wetlands Ecol Manage 13:531–51.
Rejmánek M, Lepš J (1996) Negative association can reveal interspe-
cific competition and reversal of competitive hierarchies during suc-
cession. Oikos 76:161–8.
Ripley BD (1977) Modeling spatial patterns. J R Stat Soc B 39:172–212.
Smith TJ III (1987) Seed predation in relation to tree dominance and
distribution in mangrove forests. Ecology 68:266–73.
Saenger P, Snedaker SC (1993) Pantropical trends in mangrove above-
ground biomass and annual litterfall. Oecologia 96:293–9.
Sherman RE, Fahey TJ, Martinez P (2003) Spatial patterns of biomass
and aboveground net primary productivity in a mangrove ecosys-
tem in the Dominican Republic. Ecosystems 6:384–98.
Sheue CR, Liu HY, Yong JWH (2003) Kandelia obovata (Rhizophora-
ceae), a new mangrove species from Eastern Asia. Taxon 52:287–94.
Sousa WP, Mitchell BJ (1999) The effects of predators on plant distri-
butions: is there a general pattern in mangroves? Oikos 86:55–66.
Sousa WP, Kennedy PG, Mitchell BJ, et al. (2007) Supply-side ecology
in mangroves: do propagule dispersal and seedling establishment
explain forest structure? Ecol Monogr 77:53–76.
Sousa WP, Quek SP, Mitchell BJ (2003) Regeneration of Rhizophora
mangle in a Caribbean mangrove forest: interacting effects of can-
opy disturbance and a stem-boring beetle. Oecologia 137:436–45.
Journal of Plant Ecology
Sukardjo S, Yamada I (1992) Biomass and productivity of a Rhizophora
mucronata Lamarck plantation in Tritih, Central Java, Indonesia.
Forest Ecol Manage 49:195–209.
Tam NFY, Wong YS, Lan CY, et al. (1995) Community structure and
standing crop biomass of a mangrove forest in Futian Natural Re-
serve, Shenzhen, China. Hydrobiologia 295:193–201.
Thioulouse J, Chessel D, Dolédec S, et al. (1997) ADE-4: a multivariate
analysis and graphical display software. Stat Comput 7:75–83.
Valiela I, Brown JL, York JK (2001) Mangrove forests: one of
the world’s threatened major tropical environments. BioScience
51:807–15.
Walters BB (2003) People and mangroves in the Philippines: fifty years
of coastal environmental change. Environ Conserv 30:293–303.
Weng Y (1999) Biomass and productivity of five mangrove communi-
ties in Yingluo Bay of Guangxi. Guangxi Science 6:142–7(in Chinese
with English abstract).
Downloaded from http://jpe.oxfordjournals.org by Osbert Sun on August 19, 2010
Wolf A (2005) Fifty year record of change in tree spatial patterns
within a mixed deciduous forest. Forest Ecol Manage 215:212–23.
Zan QJ, Wang YJ, Liao WB, et al. (2001) The structure of Sonneratia
apetala + S. caseolaris–Kandelia candel mangrove plantations of
Futian, Shenzhen. For Res 14:610–5(in Chinese with English
abstract).
Zhang HD (1997) Studies on Futian Mangrove Wetland Ecosystems, Shenz-
hen. Guangdong: Guangdong Science and Technology Press (in
Chinese).
Zhang JT (2004) Quantitative Ecology. Beijing: Science Press (in
Chinese).