Document 91939

acta oecologica 33 (2008) 345–354
available at
journal homepage:
Original article
Spatial pattern of diversity in an old-growth
temperate forest in Northeastern China
Xugao Wang, Zhanqing Hao*, Ji Ye, Jian Zhang, Buhang Li, Xiaolin Yao
Institute of Applied Ecology, Chinese Academy of Science, P.O. Box 417, Shenyang 110016, China
article info
Article history:
Species diversity has attracted particular attention because of its significance for helping
Received 26 March 2007
determine present species performance and likely future community composition. The
Accepted 23 January 2008
spatial pattern of species diversity (species richness, abundance and Shannon diversity)
Published online 12 March 2008
in Changbai temperate forest in Northeastern China was studied to investigate the present
and likely causes for the formation of spatial patterns. To fulfill this goal, three aspects of
diversity were addressed: 1) changes in the relationships of the diversity variables, species
Changbai Mountain
richness, abundance and Shannon diversity, to sampling area and sampling design. The
Temperate forest
three diversity variables were found to respond to sampling area in a dissimilar way.
Species diversity
Sampling design had no significant effect on the diversity variable-area curves. The power
Spatial pattern
function, which was derived under the assumption that the forest was in equilibrium, did
not fit the observed species-area curves, indicating that the Changbai temperate forest was
probably not in equilibrium. 2) Variograms, used to examine the spatial structure of species
diversity, showed that the spatial structure of species diversity in the Changbai temperate
forest was weakly anisotropic. 3) Partitioning the variation of species diversity into spatial
and environmental factors indicated that the spatial pattern of the Changbai forest
community was unpredictable, probably because there were many undetermined processes controlling its development.
ª 2008 Elsevier Masson SAS. All rights reserved.
Species diversity usually refers to the species richness, abundance, or a combination of both, of a community, and is the
result of species interaction or community adaptation to its
environment over evolutionary time (Rice and Westoby,
1982). It has attracted particular attention, in large part because of its significance in helping determine the present
species performance, and likely future community composition. Ecologists have long sought to explain why numbers of
species can coexist at small spatial scales, and how these
species are distributed, especially in species-rich tropical
rain forests (He et al., 1996; Hubbell et al., 2001; Valencia
et al., 2004; Condit et al., 2006). Various hypotheses: niche
differentiation (Ashton, 1969), species competition (MacArthur, 1969) and disturbance (Denslow, 1987) have been
proposed as driving mechanisms to account for high
diversity. Species diversity patterns should emerge as the
consequence of any and all of these mechanisms. As a result,
studying species diversity patterns should help understand
the mechanisms that have generated the observed diversity
in the community.
* Corresponding author. Tel.: þ86 24 8397 0209; fax: þ86 24 8397 0300.
E-mail address: [email protected] (Z. Hao).
1146-609X/$ – see front matter ª 2008 Elsevier Masson SAS. All rights reserved.
acta oecologica 33 (2008) 345–354
Recent studies on diversity pattern have been concentrated on tropical rain forests where species diversity reaches particularly high levels (He et al., 1996; Hubbell et al.,
1999; de Oliveira and Mori, 1999; Condit et al., 2002, 2006).
For example, a 52 ha plot in Borneo and a 25 ha plot in
Ecuador support 1175 and 1104 tree species, respectively
(Wright, 2002). In contrast, the 4.2 106 km2 of temperate
forests in Europe, North America and Asia support only
1166 tree species (Latham and Ricklefs, 1993). In other
words, tree species diversity in just one small area in the
tropics is comparable to the diversity of the entire North
Temperate Zone. However, because species diversity of
one community often differs in composition, structure and
species attributes from another community, species
diversity in temperate forests has also been a fertile area
of research for many ecologists. For example, Latham and
Ricklefs (1993) suggested that regional effects caused species diversity of temperate forests to differ between eastern
Asia and North America. Busing and White (1997) demonstrated small-scale disturbances created by tree falls
enhance plant species diversity in an Appalachian oldgrowth temperate forest. Chen and Bradshaw (1999) suggested the importance of scale and gap-phase regeneration
in the spatial patterning of a temperate coniferous forest.
Lundholm and Larson (2003) showed the positive correlations between spatial environmental heterogeneity and
plant species diversity in southern Ontario, Canada. Getzin
et al. (2006) identified tree competition as having significant
influence on species spatial pattern in a Douglas-fir forest of
the Pacific Northwest region. Wang et al. (2006a,b) suggested that catastrophic fires decreased species abundances
dramatically and caused more fragmented spatial pattern,
whereas post-fire anthropogenic activities (harvest and reforestation) could differently influence species abundance
and distribution patterns in Northeastern China. Although
these studies have led ecologists to better understand the
species diversity of temperate forests, few studies have
focused specifically on how species diversity is distributed
spatially over a temperate forest, and to what extent the
diversity pattern is regulated by spatial and environmental
factors. Since most ecological processes are pattern-generating (Legendre and Fortin, 1989; Legendre, 1993), analyzing
the resulting spatial patterns may provide important clues
as to the processes that have generated them (Borcard
and Legendre, 1994).
The objective of this study was to study the spatial pattern of species diversity in an old-growth temperate forest
in Northeastern China based on a large-scale, intensively
sampled plot. Three groups of questions were addressed:
(1) How does species diversity change with plot size?
What is the relationship between species diversity and
area in different locations? What mathematical function
best describes the species-area relationships? (2) How is
species diversity distributed spatially in the old-growth
temperate forest? Is the distribution of diversity isotropic
or anisotropic? (3) Can the diversity pattern be predicted
by environmental and/or spatial factors? To what extent
do the environmental and the spatial factors (extracted
from the spatial coordinate data) contribute to the observed
spatial patterns?
Materials and methods
Study site
The study site is in the Changbai Nature Reserve, located
along the border of China and North Korea extending from
127 420 to 128 170 E and 41 430 to 42 260 N. The reserve, which
was first established in 1960 and is one of the largest
biosphere reserves in China, has been spared from logging
and other severe human disturbances. Furthermore, the
Changbai Nature Reserve joined the World Biosphere
Reserve Network under the UNESCO Man and the Biosphere
Programme in 1980. The reserve is about 200,000 ha in size
with an elevation ranging from 740 m to 2691 m at the
summit of Changbai Mountain. Changbai Mountain is the
highest mountain in Northeastern China and is the head
of three large rivers (the Songhua, Yalu and Tumen). The
topography of the northern slope is relatively moderate
(average slope <3%), whereas the other slopes are relatively
steep (average 10%). The area has a temperate continental
climate with long cold winters and warm summers. Annual
mean temperatures vary from 7.3 C in the lowest elevations
to 2.8 C near Sky Lake (a volcanic crater lake) on the mountaintop, and annual mean precipitation varies from 750 to
1340 mm. Even before the reserve was established, forest
harvesting and other human disturbances inside the reserve
had been minor compared to that at lower elevations, due in
part to difficulties of access. A major volcanic eruption
occurred between 1000 and 1410; while more recent eruptions in 1597 and 1668 were not destructive (Zhao, 1981;
Liu et al., 1992). Forest vegetation inside the reserve is
largely the result of natural succession (Zhao, 1981). Topographic and climatic variations result in vertical zones of
major forest types that are especially distinct along the
northern slope. The forests between 750 and 1100 m are
temperate, composed of Korean pine and hardwood species.
Common hardwood species include aspen (Poplus davidiana),
white birch (Betula platyphylla), Amur linden (Tilia amuresis),
oak (Quercus mongolica), Mono maple (Acer mono) and white
ash (Fraxinus mandshurica). Between 1100 and 1700 m, the
forest changes to an evergreen coniferous forest, dominated
by spruce (Picea koraiensis) and fir (Abies nephrolepis) with
typical characteristics of boreal forests. From 1700 to
2000 m, the vegetation is sub-alpine forest, dominated by
mountain birch (Betula ermanii) and larch (Larix olgensis).
Above 2000 m are tundra, bare rock, and the volcanic lake.
Hardwoods are located in the temperate forest zone areas
that extend 0.8 km outside the nature reserve (lower than
750 m in elevation) and where human activities have
transformed the pine-hardwood forests into those mainly
composed of hardwoods (Shao, 1996).
In 2004, a 25 ha broad-leaved Korean pine mixed forest
plot of 500 m 500 m was established in Changbai Nature
Reserve (Fig. 1a). Within the plot, all free-standing trees
and shrubs at least 1 cm in DBH were identified, tagged,
and mapped, and their geographic coordinates recorded following a standard field protocol (Condit, 1998). In the study
site, the terrain is relatively gentle. The elevation ranges
from 791.8 m to 809.5 m, and mean elevation is 801.5 m.
acta oecologica 33 (2008) 345–354
Fig. 1 – a. Contour map of the 25-ha Changbai temperate
plot. b. The sampling designs for the study of diversityarea relations commenced from 10 m 3 10 m quadrats at
five different locations (a–e), and the quadrat size was
doubled until the entire plot was covered.
The total number of living individuals in the first census
(2004) was 38902, comprising 52 species, 32 genera and 18
families. The main tree species include P. koraiensis, T. amurensis, Q. mongolica, F. mandshurica, Ulmus japonica, and Acer
mono. Unlike tropical rain forests without obvious dominant
species, there were 8 species with more than 1000 individuals, which accounted for 83.4% of all individuals in the
plot. Species-specific tree abundances in the plot ranged
from 1 (3 species: Sorbus pohuashanensis, Actinidia kolomakta,
and Rosa dovurica) to 7381 individuals of Corylus mandshurica,
the most abundant species. In addition, by virtue of the criterion that species with 1 individual per ha was considered
as rare species, there were 18 rare species, accounting for
34.6% of the total number of species in the plot. Mean stand
density was 1556 living trees ha1. Mean basal area was
43.2 m2 ha1 (Hao et al., 2008; Wang et al., in press).
Data analyses
In this paper, diversity refers to richness, abundance and the
Shannon diversity index. Richness is defined as the number
of species in the study area and abundance as the number
of all individuals. The Shannon diversity index has been
suggested by Margalef (1958) as a synthetic measure of
community structure.
Different methods were used to answer the questions
stated in the Introduction. For question 1, richness, abundance and Shannon diversity were measured in square plots
that ranged from 10 m 10 m to the complete 25-ha plot for
five starting locations shown in Fig. 1b (a, b, c, d and e). The
expected species-area curve (null model) was also computed
under the assumption that all species in the study area were
randomly distributed (Coleman et al., 1982; He and Legendre,
2002). The species-area relations were fitted by three models:
the power model, the exponential model and the logistic
model. The statistical criterion for the fit of a species-area
curve is the sum of squares of the residuals. The simplest
way to test whether models are significantly different is to
check the 95% confidence intervals of the model parameters
(Sokal and Rohlf, 1981). If there is no overlap in the confidence
intervals for corresponding parameters, then we conclude
they are significantly different. All the statistical tests of
significance and confidence intervals in this paper were
computed at the a ¼ 0.05 level.
For question 2, variogram analysis was used to detect the
spatial distribution of species diversity, because the semivariance is evaluated from the differences between pairs of
observations over predetermined distance classes and emphasizes heterogeneity (Legendre and Legendre, 1998). A typical variogram can be described using three basic parameters:
(1) the range is the distance at which the semi-variance ceases
to increase (i.e. the spatial influence disappears); (2) the sill is
the semi-variance value that the variogram reaches at the
range; in theoretical variograms, the sill equals the overall
variance of a variable; and (3) the nugget effect is the ordinate
value of the variogram at distance zero; it need not be equal to
zero. It corresponds to the local variation occurring at scales
finer than the sampling interval, such as sampling error,
fine-scale spatial variability, and measurement error. The
ratio of the nugget effect to the sill is referred to as the relative
nugget effect; it can be used to evaluate sampling error and
short-scale spatial effect. To determine the strength of anisotropy, variograms of richness, abundance and Shannon
diversity were computed in four geographic directions:
0 (south-north: SN), 90 (west-east: WE), 45 (SW to NE) and
135 (SE to NW).
For question 3, the spatial patterns of diversity in Changbai
temperate forest were explored through environmental and
spatial factors, following a polynomial trend-surface analysis
(Borcard et al., 1992). The ‘spatial’ data matrix was constructed
from all quadrat locations (x and y coordinates) in the Changbai plot, by including all terms of a cubic trend-surface
polynomial equation (the x and y geographic coordinates
were centred on their respective means before computing
the other terms of the geographic polynomial). A stepwise selection procedure was used to discard the terms of the trend
acta oecologica 33 (2008) 345–354
S ¼ b1 x þ b2 y þ b3 x þ b4 xy
A ¼ b1 y þ b2 xy2 þ b3 y3
D ¼ b1 x þ b2 x2 þ b3 x2 y
Topographical data (elevation and slope) were the only
synthetic environmental variables available; they are related
to and indicators of several abiotic factors, such as drainage
condition, nutrient flow, etc. All variables were measured at
the scale of 10 m 10 m quadrats in the 25-ha plot. The
same elimination procedure as for question 1 above was
applied to the environmental data and their combination
(relative elevation z1 and slope z2), resulted in the following
equations for species richness (S ), abundance (A), and Shannon diversity (D), respectively:
S ¼ c1 z1 þ c2 z22 þ c3 z1 z2
A ¼ c1 z1 þ c2 z31
D ¼ c1 z1 þ c2 z2
Partial regression analysis was applied to measure the
amount of variation in each of the three vectors of diversity
data in turn that could be explained by the environmental factor, spatial variable or their interactions. The total variation of
a variable is decomposed into four fractions (Borcard et al.,
1992; He et al., 1996), as described below:
(a) Pure spatial contribution. This is the pure spatial effect
that cannot be described by the environmental variables, that is, is independent of any environmental
(b) Spatial þ environmental contribution. This is the proportion of variation explained by the environmental and
the spatial variables together. Two types of situation
may be responsible for this fraction of variation: firstly,
diversity may vary spatially as a function of the
environmental factors in the model or, secondly, there
may exist other processes, unidentified in the regression model under study, which control both the
species diversity and the environmental factors in the
(c) Pure environmental contribution. This is the proportion of the
diversity variation independent of any spatial structure.
(d) Undetermined contribution. This fraction, which measures
the unexplained fraction of variation, does not possess
large-scale spatial structure which would have come out
in fractions (b) or (c). It may be the consequence of stochastic fluctuations or sampling error, or it may reflect some
spatially structured variation which exists at a scale
smaller than the sampling scale.
Species diversity
In the Changbai plot, the species-area relationship described
the tendency for species richness to increase with sampling
area; a relationship whose slope declines (but remains
positive) as sampling area increased. When the sampling
area increased to 5 ha, there were about 42 species in all
sampling designs, approximately 80% of the total number of
species in the Changbai plot, then the curve slowed becoming
asymptotic (Fig. 2). Among the three models, the logistic
model best described the species-area curves (Fig. 3), with
the lowest sums of residuals (Table 1). Furthermore, the
different sampling designs (Fig. 1b) did not significantly affect
the parameters of the model. For example, the limits of the
95% confidence intervals for parameter a in the logistic model
for sample designs a and b are (60.95, 71.51) and (50.72, 63.68),
respectively (Table 1), which shows no significant difference.
However, the expected species-area curve was significantly
different from the observed species-area curves (Fig. 2),
showing that species in the Changbai temperate forest plot
are not randomly distributed.
The relations between abundance and sampling area are
extremely well fitted by linear models (Fig. 4). The confidence
intervals of the parameters of the linear models indicate that
these models are not significantly different among sampling
designs. For example, the limits of the confidence intervals
of slope for designs a and b are (2360.23, 2391.19) and
(2377.22, 2402.12), respectively. The predicted values of
abundance would not vary significantly for different sampling
designs. The density (individuals/unit area) -area curves
(Fig. 5) show that the density in different sampling designs
varies greatly, especially within sampling areas less than
5–10 ha. This indicates that with smaller sample sizes, the
variance of the estimates would be very large.
Species richness
surface equation whose contribution to each of the three
vectors of species diversity data was not significant (P < 0.01).
The following terms were retained for the three trend surface equations of species richness (S ), abundance (A), and
Shannon diversity (D), respectively:
Area (ha)
Fig. 2 – Species richness-area curves. a–e indicate different
sampling designs of Fig. 1b. f is the expected richness-area
curve under the assumption that all species are randomly
distributed over the study area.
acta oecologica 33 (2008) 345–354
parameters for the same model show that for different sampling designs, the model parameters may not be significantly
Species richness
Logistic model
Exponential model
Power model
Area (ha)
Fig. 3 – Logistic, Exponential and Power models fitted to the
species richness-area relations for sampling design a.
The relationship between Shannon diversity and area also
shows that Shannon diversity varies greatly within small
sample areas in different sampling designs in the Changbai
temperate forest plot (Fig. 6). The Shannon diversity-area
curves are best fitted neither by the power nor by the exponential model, but by a parabolic model (Fig. 7 and Table 2).
In addition, the confidence intervals of corresponding
Spatial structure of species diversity
The variograms of richness (Fig. 8, 1a, 2a) show some evidence
of anisotropy in that the semi-variances increase relatively
quickly with increased distances for 90 and 135 , whereas
there are only slight increases for the other two directions.
However, although the semi-variances for species richness
are not equal in the four directions, the difference is not obvious, probably because of relatively low species diversity in
Changbai temperate forest. In addition, the relative nugget
effect for the four directions is similar, about 54% (Fig. 8, 1a
and 2a). The spatial structure of abundance shows that the
rapid increase in semi-variances in the small distance classes
indicates that random variation characterizes the distributions of abundance (Fig. 8, 1b and 2b). In addition, the nugget
effects for the four directions are high, more than 85%,
although the nugget effects of the 0 and 90 variograms
seem lower than for the 45 and 135 directions. In addition,
in the 135 direction, the semi-variance of abundance
decreases greatly after the distance exceeds about 200 m,
which is obviously different to the other three directions.
The spatial structure of Shannon diversity shows pure nugget
effect in the 0 and 45 direction, which exhibits no spatial
auto-correlation at the study scale. However, in the other
directions, the semi-variances increase with distance,
Table 1 – Comparisons of three species (S)-area (A) models: logistic, exponential and power. a–e indicate the different
sampling designs of Fig. 1. f is the expected species-area curve, and g is the large-tree group. ‘Residual’ is the sum of
squared residuals after fitting the given model, and ‘conf. interval’ is the half-width of the 95% confidence intervals of the
parameter values. The logistic model is the best one to fit species-area curves, whereas the power model is the worst
Models sampling
Logistic model
Exponential model
Power model
S ¼ a=b þ expðglnðAÞÞ
S ¼ a þ blnðAÞ
S ¼ aAb
Parameters conf.
a ¼ 66.23 5.28
b ¼ 1.07 0.14
g ¼ 0.48 0.05
a ¼ 57.2 6.48
b ¼ 0.81 0.18
g ¼ 0.39 0.06
a ¼ 58.81 7.31
b ¼ 0.91 0.21
g ¼ 0.42 0.07
a ¼ 44.29 6.13
b ¼ 0.57 1.96
g ¼ 0.36 0.07
a ¼ 53.46 4.94
b ¼ 0.84 0.14
g ¼ 0.29 0.06
a ¼ 103.9 1.35
b ¼ 1.81 0.03
g ¼ 0.5 0.01
a ¼ 31.38 2.8
b ¼ 1.03 0.14
g ¼ 0.6 0.07
Parameters conf.
Parameters conf.
a ¼ 32.54 0.7
b ¼ 6.28 0.32
a ¼ 31.17 1.19
b ¼ 0.18 0.02
a ¼ 32.33 0.49
b ¼ 6.17 0.22
a ¼ 30.84 0.91
b ¼ 0.17 0.01
a ¼ 31.24 0. 51
b ¼ 6.08 0.24
a ¼ 28.82 0.99
b ¼ 0.18 0.01
a ¼ 29.43 0. 61
b ¼ 6.12 0.27
a ¼ 27.67 0.84
b ¼ 0.19 0.01
a ¼ 30.22 0. 67
b ¼ 6.7 0.3
a ¼ 28.47 1.09
b ¼ 0.2 0.02
a ¼ 35.94 0.45
b ¼ 5.59 0.2
a ¼ 35.13 1.09
b ¼ 0.14 0.01
a ¼ 15.76 0.35
b ¼ 3.7 0.16
a ¼ 15 0.76
b ¼ 0.2 0.02
acta oecologica 33 (2008) 345–354
Shannon diversity index
Species abundance (N)
Area (ha)
Fig. 4 – Abundance-area curves for the different sampling
designs of Fig. 1b.
especially in the 135 direction (Fig. 8, 1c and 2c). In the study
area of the Changbai temperate forest, the spatial structure of
Shannon diversity is closer to richness than abundance at the
scales observed.
Species spatial patterns
Area (ha)
The total variance of richness is 3.07 (Fig. 8, 1a and 2a) and the
coefficient of variation (CV) is 28.5%. The explained portion of
variation (a þ b þ c) is only 7.7% of the total variation in the
richness data. However, the undetermined proportion (d) is
very high (Fig. 9), indicating that the contributions a, b and c
to the spatial patterns of diversity are very low. The spatially
Fig. 6 – Shannon diversity-area curves. a–e indicate
different sampling designs in Fig. 1b.
structured environmental contribution (b) is higher than for
the abundance data.
The total variance of the abundance data is 110.74 (Fig. 8, 1b,
2b) and the coefficient of variation (CV) is 44.5%. The results
show that the explained portion (a þ b þ c) accounts for 2.7%
of the total variation of the abundance data (Fig. 9). The topographic and spatial contribution (a, b, and c) is, similarly,
very low, indicating that the relationship of the abundance
data to topographic and spatial factors is weak. The
undetermined proportion (d) for abundance is also very
high (97.3%).
Shannon diversity index
Density (N/ha)
Parabolic model
Exponential model
Power model
Area (ha)
Fig. 5 – Density-area curves. a–e indicates different
sampling designs in Fig. 1b.
Fig. 7 – Parabolic model, exponential model and power
models fitted to the Shannon diversity-ln(area) relation for
sampling design a.
acta oecologica 33 (2008) 345–354
Table 2 – Comparison of three Shannon diversity (D)-area (A) models: parabolic, exponential and power. a to e indicate the
different sampling designs of Fig. 1. ‘Residual’ is the sum of squared residuals after fitting the given model, and ‘conf.
interval’ is the half-width of the 95% confidence intervals of the parameter values. The parabolic model is the best one to fit
Shannon diversity-area curves, whereas the power model is the worst
Models sampling
Parabolic model
D ¼ a þ blnðAÞ þ glnðAÞ
Exponential model
Power model
D ¼ a þ blnðAÞ
D ¼ aAb
Parameters conf.
Parameters conf.
a ¼ 2.17 0.04
b ¼ 0.1 0.014
g ¼ 0.02 0.006
a ¼ 2.56 0.04
b ¼ 0.04 0.012
g ¼ 0.05 0.006
a ¼ 2.01 0.047
b ¼ 0.09 0.016
g ¼ 0.01 0.007
a ¼ 2.05 0.043
b ¼ 0.13 0.015
g ¼ 0.03 0.006
a ¼ 2.01 0.018
b ¼ 0.1 0.006
g ¼ 0.01 0.003
a ¼ 2.08 0.04
b ¼ 0.09 0.02
a ¼ 2.07 0.05
b ¼ 0.04 0.01
a ¼ 2.35 0.07
b ¼ 0.03 0.03
a ¼ 2.35 0.07
b ¼ 0.01 0.014
a ¼ 1.97 0. 04
b ¼ 0.08 0.01
a ¼ 1.97 0.04
b ¼ 0.04 0.01
a ¼ 1.92 0. 05
b ¼ 0.12 0.02
a ¼ 1.92 0.06
b ¼ 0.056 0.014
a ¼ 1.98 0. 02
b ¼ 0.1 0.01
a ¼ 1.98 0.02
b ¼ 0.047 0.004
The total variance for Shannon diversity is 0.12 (Fig. 8, 1c, 2c);
its coefficient of variation (CV) is only 25.6%. However, the
results show that the explained portion (a þ b þ c) accounts
for only 6.7% of the total variation of the Shannon diversity
data (Fig. 9) of which topographic and spatial factors make
low contributions. The spatially structured environmental
contribution (b) is higher than for the abundance data but
similar to the richness data. As before, the undetermined
proportion (d) is high.
Discussion and conclusion
Species diversity
The three diversity variables change differently with increasing sampling area, because they represent two categories of
variables that have very different spatial properties. Abundance is additive when aggregated across sampling areas,
whereas richness and Shannon diversity are non-additive
(He and Legendre, 1996; Legendre and Legendre, 1998; He
et al., 2002). For example, assume n1 and n2 are the abundances of species in two adjacent subplots, and s1 and s2 are
their corresponding species richness values. The total abundance in the two combined subplots n ¼ n1 þ n2, whereas the
total number of species s s1 þ s2 (the equal sign holds only
when the two subplots have totally different species composition). Shannon diversity is a combination of species richness
and abundance, which is also non-additive. As a result, the
Shannon diversity in the combined plot also does not equal
the sum of that in the two subplots.
The species-area relationship is well fitted by the logistical
model but not the power model. The power model assumes
a dynamic equilibrium (Preston, 1960; MacArthur and Wilson,
Parameters conf.
1967). Our results suggest that the temperate forest under
study would not be in a state of equilibrium. If the forest
were in equilibrium, species abundance would be estimated
in an unbiased way by any sample size as trees would be randomly distributed throughout the plot. However, in the
Changbai plot, the density in the different designs varied
greatly, especially within small sampling areas, indicating
that the trees were not randomly distributed.
The diversity (richness, abundance and Shannon diversity)area curves may be influenced by the spatial patterns of species
distributions (Hubbell and Foster, 1983). However, in the
Changbai temperate forest, the sampling location does not
appear to significantly influence these species diversity-area
curves. With different sampling designs (Fig. 1b), the same theoretical models display no significant differences (Table 1).
However, because the samples are not independent of one another, the confidence intervals of the parameters are likely to
be narrower than they should be for the normal a ¼ 5% level
(Legendre, 1993). Therefore, only well-separated confidence intervals should lead to the conclusion that parameters differ significantly. As a result, sampling design d (Table 1) should not be
considered significantly different from designs a and b, despite
the difference between the limits of the 95% confidence intervals for parameter a in the logistic model for sampling design
d (38.16, 50.42) and the confidence limits for sampling designs
a (60.95, 71.51) and b (50.72, 63.68). However, He et al. (1996)
found similar species diversity-area curves changed significantly with different sampling locations in tropical rain forests.
Given the relative simplicity and low species diversity compared to tropical rain forests, the results observed in Changbai
temperate forests are not surprising.
The spatial structure of species diversity
Variograms of richness, abundance and Shannon diversity
in the Changbai temperate forests showed differences
acta oecologica 33 (2008) 345–354
Shannon diversity
Shannon diversity
Distance (m)
Fig. 8 – Variograms of richness (1a–2a), abundance (1b–2b) and Shannon diversity (1c–2c) in the four geographic directions:
08 is east (E)-west (W), 908 is south (S)-north (N), 458 is SE-NW and 1358 is NE-SW. The horizontal lines indicate the overall
variance of the variables in the plot.
between the four directions, but are not clearly anisotropic,
unlike that found in tropical forests. For example, He et al.
(1996) studied the spatial pattern of diversity in a rainforest
in Malaysia, and found that the spatial structure of diversity
was clearly anisotropic. Furthermore, in this Changbai
study, all variograms of the three diversity indices showed
relatively high nugget effects, probably because of smallscale processes that may operate in the temperate forest.
Some interesting spatial features may be detected at finer
scales than the smallest scale used here (¼10 m). These
small-scale processes may include competition, predation,
dispersal, microbial interactions, etc., which could result
in the observed spatial heterogeneity of diversity in the
Changbai temperate forests. For example, some small-scale
disturbances, such as windthrow, fire and insects, are
known to promote the regeneration of a diverse array of
species. These small-scale disturbances create open places
favorable for some pioneering plants, such as white birch
and aspen, which might result in different species diversity
compared with that in later successional stages (Hao et al.,
1994, 2002; Wu, 1998). Also, the intensity of the disturbance
could lead to different species diversities. The Intermediate
Disturbance Hypothesis states that higher species diversity
occurs at intermediate levels of disturbance because species
coexistence is maintained at a non-equilibrium state and no
strong competitor can dominate completely (Connell, 1978).
However, due to its cold climate and relatively gentle
terrain, species richness in the Changbai plot is low, which
might cause the spatial structure of diversity to be weakly
anisotropic in these temperate forests.
acta oecologica 33 (2008) 345–354
but this is certainly not the case here since the survey has
been exhaustive. Finally, niche differentiation, species specificity and the lack of dominant controlling forces (many processes controlling the structure of temperate communities,
each one playing but a small role) may be invoked.
Shannon diversity
a: pure space
c: pure environment
b: space + environment
d: undetermined
This paper is sponsored by the Knowledge Innovation Program of the Chinese Academy of Sciences (KZCX2-YW-430),
National Natural Science Foundation of China (30700093 and
30570306), and National Key Technologies R&D Program of
China (2006BAD03A09). The authors thank all those who
provided helpful suggestions and critical comments on this
manuscript, including Fangliang He, He Hong S, Michael
Papaik, and Bill Loneragan.
Fig. 9 – Relative percentage of variation partitioning of
species richness, abundance and Shannon diversity.
Spatial patterns and controlling processes
Partitioning the variation in species diversity helps us to
understand the community structure in Changbai temperate
forest, and the processes that may have contributed to its
formation. In this study there are several common features
to the partitioning of variation in the richness, abundance
and Shannon diversity data, which indicate similar underlying controlling processes in the Changbai forest, with fairly
large pure spatial components (a) and very small pure environmental components (c), as well as extremely high undetermined proportions (d). Relatively high (a) may result either
from spatially structured environmental factors, from
spatially structured historical processes, or from environment
independent processes, such as growth, reproduction, predation and competition with neighbours acting to shape the
community. Low environmental explanation (c) may be attributed, on the one hand, to the relatively flat topography of the
study area which is typical of Changbai temperate forests; on
the other hand, it may result from the absence of dominant
environmental controlling factors in this study area. Whatever the cause, clearly the observed diversity does not vary
as a function of topography. According to our hypothesis, if
the community is under equilibrium, then the predictable
proportion (a, b and c) should be high and the undetermined
component (d) low. In other words, if the diversity varies as
a function of the environmental variables, the amount of
explained variation in fraction (a þ b) is expected to be high
and significant; and if the diversity varies as a function of
the spatial variables, then fraction (b þ c) is expected to be
significantly high. The high unexplained components (d) in
this study seem to be indicating the temperate forest in
Changbai Nature Reserve is not in equilibrium. There are
some possible explanations for the high unexplained proportions (d). One is that there is only a small amount of variability
to be explained in this temperate forest, and this could not be
captured by the trend-surface equations at the small-scales
studied. Another possible explanation is high sampling error,
Ashton, P.S., 1969. Speciation among tropical forest trees: some
deductions in the light of recent evidence. In: LoweMcConnell, R.H. (Ed.), Speciation in Tropical Environment.
Academic Press, London, pp. 155–196.
Borcard, D., Legendre, P., 1994. Environmental control and spatial
structure in ecological communities: an example using
Oribatid mites (Acari, Oribatei). Environmental and Ecological
Statistics 1, 37–53.
Borcard, D., Legendre, P., Drapeau, P., 1992. Partialling out the spatial
component of ecological variation. Ecology 73, 1045–1055.
Busing, R.T., White, P.S., 1997. Species diversity and small-scale
disturbance in an old-growth temperate forest: a consideration
of gap partitioning concepts. Oikos 78, 562–568.
Chen, J., Bradshaw, G.A., 1999. Forest structure in space: a case
study of an old growth spruce-fir forest in Changbaishan
Natural Reserve, PR China. Forest Ecology and Management
120, 219–233.
Coleman, B.D., Mares, M.A., Willig, M.R., Hsieh, Y.H., 1982.
Randomness, area, and species richness. Ecology 63, 1121–1133.
Condit, R., 1998. Tropical Forest Census Plots: Methods and
Results from Barro Colorado Island, Panama and
a Comparison with Other Plots. Springer, Berlin.
Condit, R., Pitman, N., Leigh, E.G., Chave, J., Terborgh, J., Foster, R.
B., Nunez, P., Aguilar, S., Valencia, R., Villa, G., MullerLandau, H.C., Losos, E., Hubbell, S.P., 2002. Beta-diversity in
tropical forest trees. Science 295, 549–562.
Condit, R., Ashton, P., Bunyavejchewin, S., Dattaraja, H.S.,
Davies, S., Esufali, S., Ewango, C., Foster, R., Gunatilleke, I.A.U.
N., Gunatilleke, C.V.S., Hall, P., Harms, K.E., Hart, T.,
Hernandez, C., Hubbell, S., Itoh, A., Kiratiprayoon, S.,
LaFrankie, J., Lao, S., Makana, J.-R., Noor, Supardi, Md, N.,
Rahman Kassim, A., Russo, S., Sukumar, R., Samper, C.,
Suresh, H.S., Tan, S., Thomas, S., Valencia, R., Vallejo, M.,
Villa, G., Zillio, T., 2006. The importance of demographic
niches to tree diversity. Science 313, 98–101.
Connell, J.H., 1978. Diversity in tropical rain forest and coral reefs.
Science 199, 1304–1310.
Denslow, J.S., 1987. Tropical rain forest gaps and tree species
diversity. Annual review of ecology and Systematics 18, 431–451.
Getzin, S., Dean, C., He, F., Trofymow, J.A., Wiegand, K.,
Wiegand, T., 2006. Spatial patterns and competition of tree
species in a Douglas-fir chronosequence on Vancouver Island.
Ecography 29 (5), 671–682.
acta oecologica 33 (2008) 345–354
Hao, Z., Zhao, S., Tao, D., 1994. Species diversity and its seasonal
dynamics of herbs in a broad-leaved Korean pine forest on the
northern slope of Changbai Mountain. Biodiversity Science 2
(3), 125–132 (in Chinese).
Hao, Z., Guo, S., Cao, T., 2002. Species diversity and distribution
patterns in Changbai Mountain. Liaoning Science and
Technology Publishing House, Shenyang (in Chinese).
Hao, Z., Li, B., Zhang, J., Wang, X., Ye, J., Yao, X., 2008. Broad-leaved
Korean pine (Pinus koraiensis) mixed forest plot in Changbaishan
(CBS) of China: Community composition and structure. Acta
Phytoecologica Sinica 32 (2), 238–250 (in Chinese).
He, F., Legendre, P., 1996. On species-area relations. American
Naturalist 148, 719–737.
He, F., Legendre, P., 2002. Species diversity patterns derived from
species-area models. Ecology 83, 1185–1198.
He, F., Legendre, P., LaFrankie, J.V., 1996. Spatial pattern of
diversity in a tropical rain forest in Malaysia. Journal of
Biogeography 23, 57–74.
He, F., LaFrankie, J.V., Song, B., 2002. Scale dependence of tree
abundance and richness in a tropical rain forest, Malaysia.
Landscape Ecology 17, 559–568.
Hubbell, S.P., Foster, R.B., 1983. Diversity of canopy trees in
Neotropical forest and implications for conservation. In:
Sutton, S.L., Whitmore, T., Chadwick, A.C. (Eds.), Tropical Rain
Forest: Ecology and Management. Blackwell Science, Oxford,
pp. 25–41.
Hubbell, S.P., Foster, R.B., O’Brien, S.T., Harms, K.E., Condit, R.,
Wechsler, B., Wright, S.J., Loo de Lao, S., 1999. Light-gap
disturbance, recruitment limitation, and tree diversity in
a Neotropical forest. Science 283, 68–71.
Hubbell, S.P., Ahumada, J.A., Condit, R., Foster, R.B., 2001. Local
neighborhood effects on long-term survival of individual trees
in a neotropical forest. Ecological Research 16, 859–875.
Latham, R.E., Ricklefs, R.E., 1993. Continental comparisons of
temperate-zone tree species diversity. In: Ricklefs, R.E.,
Schluter, D. (Eds.), Species Diversity in Ecological Communities.
University of Chicago Press, Chicago, pp. 294–314.
Legendre, P., 1993. Spatial autocorrelation: trouble or new
paradigm? Ecology 74, 1659–1673.
Legendre, P., Fortin, M.J., 1989. Spatial pattern and ecological
analysis. Vegetatio 80, 107–138.
Legendre, P., Legendre, L., 1998. Numerical Ecology, second ed.
Elsevier Science, Amsterdam, The Netherlands.
Liu, Q., Wang, Z., Wang, S., 1992. Recent volcano eruptions and
vegetation history of alpine and sub-alpine of Changbai
Mountain. Forest Ecosystem Research 6, 57–62 (in Chinese).
Lundholm, J.T., Larson, D.W., 2003. Relationships between spatial
environmental heterogeneity and plant species diversity on
a limestone pavement. Ecography 26, 715–722.
MacArthur, R.H., 1969. Patterns of communities in the tropics.
Biological Journal of the Linnean Society 1, 19–30.
MacArthur, R.H., Wilson, E.O., 1967. The Theory of Island
Biogeography. Princeton Univ. Press, Princeton, N.J.
Margalef, R., 1958. Information theory in ecology. General
Systems 3, 36–71.
de Oliveira, A.A., Mori, S.A., 1999. A central Amazonian terra firme
forest. I. High tree species richness on poor soils. Biodiversity
and Conservation 8, 1219–1244.
Preston, F.W., 1960. Time and space and the variation of species.
Ecology 41, 611–627.
Rice, B., Westoby, M., 1982. Plant species richness at the 0.1
hectare scale in Australian vegetation compared to other
continents. Vegetatio 52, 129–140.
Shao, G., 1996. Potential impacts of climate change on a mixed
broadleaved-Korean pine forest stand: a gap model approach.
Climatic Change 34, 263–268.
Sokal, R.R., Rohlf, F.J., 1981. Biometry. Freeman, San Francisco.
Valencia, R., Foster, R.B., Villa, G., Condit, R., Svenning, J.C.,
Hernandez, C., Romoleroux, K., Losos, E., Magard, E.,
Balslev, H., 2004. Tree species distributions and local habitat
variation in Amazon: large forest plot in eastern Ecuador.
Journal of Ecology 92, 214–229.
Wang, X., He, H.S., Li, X., Hu, Y., 2006a. Assessing the cumulative
effects of post-fire management on forest landscape dynamics
in northeastern China. Canadian Journal of Forest research 36,
Wang, X., He, H.S., Li, X., Chang, Y., Hu, Y., Xu, C., Bu, R., Xie, F.,
2006b. Simulating the effects of reforestation on a large
catastrophic fire landscape in Northeastern China. Forest
Ecology and Management 225, 82–96.
Wang, X., Hao, Z., Ye, J., Zhang, J., Li, B., Yao, X., in press. Spatial
variation of species diversity across scales in an old-growth
temperate forest of China. Ecological Research, in press. doi:
Wright, S.J., 2002. Plant diversity in tropical forests: a review of
mechanisms of species coexistence. Oecologia 130, 1–14.
Wu, G., 1998. Regeneration dynamics of tree species in gaps of
Korean pine broad-leaved mixed forest in Changbai Mountains.
Chinese Journal of Applied Ecology 9 (5), 449–452 (in Chinese).
Zhao, D., 1981. Preliminary study of the effects of volcano on
vegetation development and succession. Forest Ecosystem
Research 2, 81–87 (in Chinese).