SILVA FENNICA
Silva Fennica 46(4) research articles
www.metla.fi/silvafennica · ISSN 0037-5330
The Finnish Society of Forest Science · The Finnish Forest Research Institute
Species-Habitat Associations in a
Northern Temperate Forest in China
Chunyu Zhang, Yazhou Zhao, Xiuhai Zhao and Klaus von Gadow
Zhang, C., Zhao, Y., Zhao, X. & Gadow, K. v. 2012. Species-habitat associations in a northern
temperate forest in China. Silva Fennica 46(4): 501–519.
This contribution identiies species-habitat associations in a temperate forest in north-eastern
China, based on the assumption that habitats are spatially autocorrelated and species are
spatially aggregated due to limited seed dispersal. The empirical observations were obtained
in a large permanent experimental area covering 660 × 320 m. The experimental area was
subdivided into four habitat types using multivariate regression tree (MRT) analysis. According to an indicator species analysis, 38 of the 47 studied species were found to be signiicant
indicators of the MRT habitat types. The relationships between species richness and topographic variables were found to be scale-dependent, while the great majority of the species
shows distinct habitat-dependence. There are 188 potential species-habitat associations,
and 114 of these were signiicantly positive or negative based on habitat randomization. We
identiied 139 signiicant associations using a species randomization. A habitat is not a closed
system it may be both, either a sink or a source. Therefore, additional to the randomization,
the Poisson Cluster Model (PCM) was applied. PCM considers the spatial autocorrelation of
species and habitats, and thus appears to be more realistic than the traditional randomization
processes. It identiied only 37 associations that were signiicant. In conclusion, the deviation
from the random process, i.e. the high degree of species spatial mingling may be explained
by persistent immigration across habitats.
Keywords dispersal limitations, indicator species, spatial autocorrelation, species richness,
topographic differentiation
Addresses Zhang and X. Zhao, Key Laboratory for Forest Resources & Ecosystem Processes
of Beijing, Beijing Forestry University, Beijing 100083, China; Y. Zhao, Department of
Landscape Architecture, School of Architecture, Tsinghua University, Beijing 100084, China;
Gadow, Faculty of Forestry and Forest Ecology, Georg-August-University Göttingen, Büsgenweg 5, D-37077 Göttingen, Germany E-mail (Zhang) zcy_0520@163.com
Received 1 March 2012 Revised 10 September 2012 Accepted 12 September 2012
Available at http://www.metla.i/silvafennica/full/sf46/sf464501.pdf
501
Silva Fennica 46(4), 2012
1 Introduction
Spatial distributions of forest trees often exhibit
patterns correlating with the variation of soil
chemistry or topography in tropical forests
(Harms et al. 2001, Itoh et al. 2003, Russo et al.
2005, Yamada et al. 2006, 2007, John et al. 2007)
and in temperate forests (Zhang et al. 2009, 2010).
This suggests that the ecological organization
caused by niche differentiation may be important for maintaining species diversity and species coexistence. If environmentally biased spatial
distributions principally result from niche differentiation, plant species should show particular
habitat preferences. They would preferably occur
in localities where they have competitive advantages, although spatial autocorrelation cannot be
ignored when considering species-habitat associations (Legendre and Legendre 1998).
A common assumption of most traditional statistical methods for species-habitat associations
is that individuals are independently distributed
with respect to conspeciics (Condit 1996, Clark
et al. 1998, Plotkin et al. 2000). But the independence assumption is often violated by the patterns
produced by short-distance dispersal and recruitment processes. The limited dispersal of seeds
and short-distance recruitments would contribute
to the spatial autocorrelation of species distributions (Condit 1996, Clark et al. 1998, Plotkin et
al. 2000). Thus, the assumptions of independence
of sample units are often violated by the pattern
caused by the dispersal limitations and dependent
recruitment processes of trees and shrubs.
To test the contribution of habitat specialization
to species coexistence, the relationships between
the species spatial distribution and environmental
factors need to be studied. In the northern temperate forests of China, the distribution patterns of
individuals within a plant population generally
tend to be more aggregated than random (Zhang
et al. 2009). Furthermore, signiicant correlations
between species and soil nutrients were found in
these forests (Zhang et al. 2010). This suggests
that habitat preferences are potentially important
in explaining the spatial variation in tree communities. Nakashizuka (2001) maintained that habitat specialization remains a prominent hypothesis
to explain the species coexistence in a temperate
forest community.
502
research articles
The objective of this study is to analyse some
of the mechanisms generating differences in species abundance across habitat types. The fully
mapped experimental area of 21 ha is located in
a multi-species forest ecosystem in North-Eastern
China. We assume that habitats are spatially autocorrelated and that the range of seed dispersal is
limited. Based on previous ield observations, we
expect substantial species-habitat associations in
the experimental area. Speciic objectives of this
study are (1) to determine possible scale-dependent associations between species richness and
topographic variables; (2) to identify indicator
species for a particular habitat and (3) to examine
possible associations of trees and shrubs with
distinct habitats. We will also discuss the effect
of habitat differentiation in maintaining a high
species diversity in the Jiaohe temperate forest.
2 Materials and Methods
2.1 Study Area
This study is based on a dataset obtained in a large
permanent ield plot. The experimental site is located at (43°57.897´ ~ 43°58.263´N, 127°42.789´ ~
127°43.310´E) in the Jiaohe Management Bureau
of the Forest Experimental Zone in Jilin province,
in Northeastern China. The research plot measures 320 m × 660 m and covers an area of 21.12
hectares. The altitude in the experimental area
ranges from 425.3 m to 525.8 m above sea level.
In the study area, the average annual temperature
is 3.8 °C. And the hottest month is July with an
average day temperature of 21.7 °C. The coldest
month is January with an average day temperature
of –18.6 °C. The average annual precipitation is
695.9 mm. The soil is a brown forest soil with a
rootable depth ranging between 20 and 100 cm.
The last recorded tree felling activities took place
50 years ago. The vegetation type represents a
mixed broadleaf-conifer forest with 63 species
(including three climber species).
Altogether 53 916 individual trees with a breast
height diameter (dbh) exceeding 1cm were tagged
and mapped, and their species was identiied.
The dbh value was measured at 1.3 m above
ground level. Among the 63 woody species in
Zhang et al.
the research plot there are 47 abundant species,
comprising at least one individual/ha. The species were identiied according to the records in
the Chinese Virtual Herbarium (see http://www.
cvh.org.cn/cms/).
The dominant tree species are Ulmus davidiana
var. japonica (Rehder) Nakai, Pinus koraiensis
Siebold & Zucc., Juglans mandshurica Maxim.,
Tilia mandschurica Rupr.er Maxim., Carpinus
cordata Bl., Acer mono Maxim., Fraxinus mandshurica Rupr., Tilia amurense Rupr. and Ulmus
laciniata (Trautv.) Mayr. The top ive species in
stem density are Acer mandshuricum Maxim.,
Syringa reticulata var. amurensis (Rupr.), Ulmus
davidiana var. japonica, Carpinus cordata and
Acer mono, respectively. The total basal area of
dominant tree species and stem density of the
top ive species are shown in Appendix 1 and 2.
2.2 Relationships between Species Richness
and Topography
The relative heights at the four corner nodes of
each 20 m × 20 m cell were used to develop a
variogram model of the entire research area. To
examine the association between species richness
and topographic variables at different scales, the
altitude values were estimated for different cell sizes
(5 m × 5 m, 10 m × 10 m, 30 m × 30 m, 40 m × 40 m
and 50 m × 50 m) using block kriging (Legendre
and Legendre 1998). Species richness in each cell
was counted at each of these six different scales.
Spearman rank correlation coeficients were calculated to test the relationships between species
richness and the topographic variables at each of
the six spatial scales. When the plot was subdivided
into equally dimensioned cells, the intersections of
grid lines were called “nodes”. The relative height
differences among the nodes and the elevation
of the starting node were measured. Thus it was
possible to calculate the elevation of other nodes
according to the height difference among nodes
and the measured elevation of the starting node.
The present study mainly focuses on the results
of the 20 m × 20 m cell analysis.
The elevation of a particular cell was calculated
as the mean of the elevations of its four corner
nodes. The cell slope for each of the ive cell sizes
was estimated as the mean angular deviation from
Species-Habitat Associations in a Northern Temperate Forest in China
the horizontal plane of each of the four triangular
planes which were formed by connecting three
of its adjacent corners (Harms et al. 2001). The
convexity of a cell was calculated as the elevation of the focal cell minus the mean elevation of
the eight surrounding cells (cf. Yamakura et al.
1995). For the edge cells, convexity was taken as
the elevation of the center point minus the mean of
the four corners. Positive and negative convexity
values respectively indicate convex (ridge) and
concave (valley) land surfaces. The aspect of a
cell can be obtained from the average angle of the
four triangular planes that deviate from the north
direction. Four maps show the spatial pattern of
the four topographic variables using 20 m × 20 m
cells (Fig. 1). Each cell shows the altitude (ranging from 425.3 m to 525.8 m above sea level with
100.5 m difference in altitude between the highest
and lowest cells), the convexity (ranging from –6.6 m
to 4.7 m), the slope (ranging from 1.4° to 39.2°)
and the aspect (ranging from 41.9° to 329.7°).
2.3 Habitat Classiication and Indicator
Species
Multivariate regression tree (MRT) analysis was
performed, following De’ath (2002), to classify
habitat types according to topographic conditions
and species composition. Distance-based MRT
is a relatively new statistical technique that can
be used to describe relationships between multispecies data and environmental characteristics.
The dissimilarities used in distance-based MRT
are calculated by Euclidean distances. Thus, one
obtains clusters of sites by repeated splitting of
the data, which are chosen to minimize the dissimilarity of sites within clusters.
Habitats were delineated using threshold values
for the topographic variables, while the species
data was used to ind the best thresholds. Indicator
values and associated probabilities were computed to identify the statistically signiicant indicator species in a speciic habitat type. Indicator
species analysis combines a species relative abundance with its relative frequency of occurrence
in the various groups of sites. Indicator value
(di,c) of species was calculated as the product of
the relative frequency (fi,c) and relative average
abundance (ai,c) in clusters.
503
Silva Fennica 46(4), 2012
research articles
a) Altitude
b) Slope
c) Convexity
d) Aspect
Fig. 1. Maps depicting four topographic variables at the scale of 20 × 20 m cells. a) Altitude from 425.3 m (white)
to 525.8 m (black) above sea level; b) Convexity from –6.6 m (white) to 4.7 m (black); c) Slope from 1.4°
(white) to 39.2° (black); d) Aspect from 41.9° (white) to 329.7° (black).
For cluster c in set K:
fi ,c =
ai ,c =
∑ j ∈c pi , j
nc
(∑ j ∈cxi , j ) / nc
K
∑ k =1((∑ j ∈ k xi , j ) / nk )
(1)
di ,c = fi ,c × ai ,c
where
pi,j = presence/absence (1/0) of species i in sample j
xi,j = abundance of species i in sample j
nc = number of samples in cluster c
The species indicator value is a maximum when
all individuals of a species are found in a single
group of sites and when the species occurs in all
sites of that group. The statistical signiicance of
the species indicator values is evaluated using a
randomization procedure (Dufrêne and Legendre
1997). All calculations were done using the R
statistical software (R Development Core Team
504
2010). MRT analysis was implemented using
the “mvpart” library of R (De’ath 2010). Indicator species analysis was performed using R’s
“labdsv” library (Roberts 2010).
Rare species with less than 50 individuals
within the 21.12 ha study area were excluded
from the species–habitat association analysis.
Trees at different life stages may have different
ecological habitat preferences, as reported by
Webb and Peart (2000), Comita et al. (2007) and
Lai et al. (2009). In this study, we only focus on
the relationship between species types and habitat
types. Thus we assume that all individuals of a
given species respond similarly to a speciic habitat type, regardless of their stage of development.
2.4 Testing Species-Habitat Associations
Most methods for testing species-habitat associations assume that trees and shrubs are independently distributed with regard to conspeciic
individuals. However, the assumption of inde-
Zhang et al.
pendence is often violated because of the limited
range of seed dispersal and recruitment (Harms et
al. 2001). In this study, we are using three methods to test this assumption. This study mainly
focuses on the results based on a Poisson Cluster
Model (PCM) analysis.
2.4.1 Randomized Habitat Maps
This section presents the case where randomized
habitat processes were modeled with dispersal
limitations of species but no spatially autocorrelated habitat features. Random habitat processes
were used to simulate habitat maps that were not
autocorrelated. The true species map was used
to indicate the dispersal limitations of species.
Spatial dependency within species was evaluated by generating a series of random habitat
maps. In these simulated maps the non-overlapping areas are identical in extent to the four
habitat types of the true maps. Each simulated
map included exactly 248 cells (20 m × 20 m) of
habitat type 1, 85 cells of habitat type 2, 52 of
habitat type 3 and 143 of habitat type 4. The
habitat types were randomly permuted among
the 528 cells.
To assess the species associations, each simulated habitat map was matched with the true tree
distribution map. Then the relative stem density
of the focal species in each habitat type was counted. This procedure was repeated 10 000 times.
Thus, the frequency distributions of the stem
density estimates for each species in each habitat
type were obtained from 10 000 simulated habitat
maps. We then compared the relative stem density
of a particular species calculated from the true
habitat map with that from the simulated habitat
maps. If the proportion of instances where {stem
density of simulated habitat maps < stem density
of true habitat map} was greater than 0.975, we
assumed that the given species was positively
associated with a particular habitat at the 0.05
level in a two-tailed test. Alternatively, if the
proportion of instances where {stem density of
simulated habitat maps > stem density of true
habitat map} was greater than 0.975, we assumed
that the given species was negatively associated
with a particular habitat at the 0.05 level in the
two-tailed test.
Species-Habitat Associations in a Northern Temperate Forest in China
2.4.2 Randomized Species Maps
This section presents the case where randomized
species processes were modeled with no dispersal
limitations of species but spatially autocorrelated
habitat features. A complete spatial randomness
(CSR) process was used to simulate no dispersal
limitations of species. The true habitat map was
used to indicate spatially autocorrelated habitats.
Spatial dependency within habitat types was
tested by generating a series of maps with random
locations for each species, using the CSR process
in the simulated species maps. Matching the true
habitat map with the simulated species maps, we
calculate the stem density of each species in each
habitat type. This procedure was repeated 10,000
times to establish the frequency distribution of the
estimated stem density. Signiicant deviations of
the observed density values from the expected
ones were assessed at 0.05 levels using a twotailed test.
2.4.3 Poisson Cluster Process
Using poisson cluster processes, dispersal limitations of species and spatial autocorrelation of
habitat maps were modeled. The Poisson Cluster
Model (PCM) was used to simulate the dispersal
limitations of species. The true habitat map was
used to indicate spatially autocorrelated habitats.
In a subsequent analysis, using Ripley’s K function, it was found that 47 species (Appendix 3)
were not randomly distributed, but signiicantly
aggregated. By choosing the PCM approach, it was
implied that clusters arise from local propagation.
The PCM models aggregation caused by local
seed dispersal or gap recruitment. Thus, it was
decided to simulate the spatial distribution of each
of these species using the Poisson cluster process.
The observed spatial aggregation is then used to
develop expected species–habitat associations.
Subsequently, we model the spatial aggregation
of the species distributions using the PCM with
Ripley’s K value as guiding parameter, as recommend by Diggle (1983), Plotkin et al. (2000) and
John et al. (2007). The PCM model can be used to
capture small-scale spatial aggregation in species
distributions that are due to aggregated dispersal
(Potts et al., 2004). The deinition and parameter
505
Silva Fennica 46(4), 2012
research articles
estimation of the Poisson cluster process follows
Plotkin et al. (2000).
3 Results
aspect (east has low values, west has high values)
is negative and highly signiicant at ine scales.
This means that the number of species is increasing when the aspect changes at close range from
west to east.
3.1 Species Richness and Topography
3.2 Habitat Types and Indicator Species
The correlation between species richness and
topography shows a distinct scale-dependence
(Table 1). Highly signiicant or signiicant negative associations between species richness and
altitude were found at iner scales, ranging from
5 m to 30 m, but vanished at higher scales. This
result may not be very surprising, considering
that fewer species will be able to establish themselves on the rather exposed higher altitudes. The
correlations are very weak although signiicant.
It means that there was almost no linear relationship between the studied habitat characteristics
and species richness. Thus, some other factors
than the studied ones may explain better the species richness or the species richness is affected
by many factors each having a low impact alone.
Convexity, which expresses the relative altitudinal difference between the focal cell and its
surrounding neighbors, does not seem to affect
species richness. Signiicantly positive associations between species richness and slope were
found at ine and coarse scales. The number of
species increases with increasing terrain steepness, but only at very close range, which means
that the increasing richness is found at the transitions from the hill bottom (or the plateau) to
the adjacent slope. This is plausible, but we are
unable to provide an explanation for the signiicant association at the 50 m scale.
The association between species richness and
The experimental area was subdivided into four
habitat types using a MRT method. The tree
size was selected using a cross-validation procedure, with the four-leaf tree clearly identiied
as having the smallest cross-validated relative
error (CV error = 0.635; see Appendix 4). The
geographical proile of the four habitat types is
shown (Fig. 2). Each 20 m × 20 m cell is assigned to one speciic habitat, as indicated by the
numbers 1, 2, 3 and 4.
The topographic attributes of the four habitat
types are presented in Table 2. Habitat type 1,
which occurs in 248 cells, occupies the lowaltitudes. It is separated from habitat type 2, 3 and
4 by the lower altitudinal boundary of 453.6 m.
Habitat type 2 (n = 85) occupies the east-facing
cells with aspects less than 187°, in the lower
right and upper left of the plot. Habitat type 3
(n = 52) and habitat type 4 (n = 143) are found in
the westward-facing cells with aspects exceeding
187°, in the upper right and upper left of the plot.
Habitat type 3 occupies altitudes below, habitat
type 4 altitudes above 465.7 m.
According to the MRT analysis, 38 of the 47
species occurring in the experimental area, were
found to be signiicant indicators of the habitat
types. Fifteen of these, listed in Appendix 5, are
signiicant indicators of habitat type 1, ive of
habitat type 2, seven of habitat type 3, and eleven
of habitat type 4.
Table 1. Correlation coeficients showing the degree of correlation between species richness
and four topographic variables at six different spatial scales
Scales
5 m×5 m
10 m × 10 m
20 m × 20 m
30 m × 30 m
40 m × 40 m
50 m × 50 m
506
Altitude
–0.1263***
–0.2537***
–0.2896***
–0.1767**
–0.1341
0.0565
Convexity
Slope
0.0056
–0.0163
–0.0934*
–0.0596
–0.2208*
–0.1869
0.0332**
–0.0084
–0.0055
0.0290
0.1966*
0.3020**
Aspect
–0.0453***
–0.0690*
–0.0542
0.0715
0.1483
0.2518*
Zhang et al.
Species-Habitat Associations in a Northern Temperate Forest in China
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
3
3
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
3
3
3
3
1
1
1
1
1
1
1
1
1
1
1
2
3
3
3
3
3
3
1
1
1
1
1
1
1
1
1
3
3
3
3
3
3
3
3
3
1
1
1
1
1
1
1
3
3
3
3
3
3
3
3
3
3
1
1
1
1
1
1
4
4
4
4
4
4
3
3
3
3
1
1
1
1
1
1
4
4
4
4
4
4
4
4
4
3
3
1
1
1
1
1
4
4
4
4
4
4
4
4
4
4
3
1
1
1
1
1
4 4
4 4
4 4
4 4
4 4
4 4
4 4
4 4
4 4
4 4
44 4
3753
1 3
1 1
1 1
1 1
4 5410 4 4
4 4 4 4
50
40 4 4 4
4 4 4 4
4 4 4 4
4 4 4 4
4 4 4 4
4 4 4 4
4 4 4 4
4 4 4 4
4 4 4 4
4 4 4 4
3 4 4 490
4
14730 3 4
1 1 1 4615
1 1 1 1
4 4
4 4
4 4
4 4
4 4
4 4
4 4
4 4
4 4
4 4
45025
2 2
2 485
2
2 2
2 1
1 1
5
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
51
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
5
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
49
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
0
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
48
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
0
4 4 4840 2 4 2 2
4 4 2 2 475
2 2 2
4 4 2 2 2 2 2
470
4 4 2 2 2 2 2
3 465
2 2 2 2 2 2
3 2 2 2 2 2 2
3 460
2 2 2 2 2 1
3 455
2 2 2 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
445
1 1 10 1 1 1 1
1 1 414 1 1 14301
1 1 1 1 1 1 1
435
1 1 1 1 1 1 1
45
2
4
4
4
3
3
3
3
1
1
1
1
1
1
1
1
4
4
4
4
4
4
4
4
4
4
2
2
2
2
1
1
Fig. 2. Map showing the distribution of the four habitat types at the 20 m × 20 m scale.
The lines show the elevation contours at 5 m intervals.
Fig. 3. Map showing the spatial distribution of two species with distinct habitat
preference: Ulmus davidiana var. japonica (red dots) and Carpinus cordata
(black dots). Background colors: green = habitat type 1; yellow = habitat type 2;
red = habitat type 3; blue = habitat type 4.
Table 2. Topographic attributes of the four habitat types.
Habitat types
Min
Habitat 1
Habitat 2
Habitat 3
Habitat 4
426.3
453.7
454.0
465.7
Elevation
Mean
439.9
466.6
459.6
489.4
Max
453.5
501.2
465.7
519.7
Min
Convexity
Mean
Max
Min
Slope
Mean
Max
Min
Aspect
Mean
Max
–6.6
–2.9
–1.4
–1.1
–0.5
0.5
0.0
0.6
1.5
4.3
2.9
4.7
1.4
5.3
5.5
3.9
10.3
15.7
13.5
15.9
38.3
39.2
33.1
37.3
41.9
140.5
188.9
187.6
187.5
164.1
231.2
239.9
329.7
185.9
275.7
286.3
Some species show distinct habitat-dependence, which becomes apparent when their spatial
distributions are mapped. An example involving
two species Ulmus davidiana var. japonica and
Carpinus cordata, each of which has a distinct
habitat preference, is presented (Fig. 3). Ulmus
davidiana var. japonica correlates negatively with
the elevation, the slope, and the convexity. Carpinus cordata, on the other hand, correlates positively with the elevation, the slope, the convexity
507
Silva Fennica 46(4), 2012
research articles
Table 3. Signiicant habitat associations, based on three
methods: randomized habitats, randomized species
and PCM. The symbol “+” represents signiicant
positive, “−”signiicant negative associations.
Habitat association
Habitat 1 +
Habitat 2 +
Habitat 3 +
Habitat 4 +
Total positive
Habitat 1 −
Habitat 2 −
Habitat 3 −
Habitat 4 −
Total negative
Grand Total
Randomized
habitats
16
2
0
24
42
18
24
25
5
72
114
Randomized
species
21
10
13
20
64
21
27
7
20
75
139
PCM
16
4
5
10
35
1
1
0
0
2
37
and the aspect. Both correlations are signiicant.
This explains the different species-habitat associations of the two species and their distribution
in the experimental area.
3.3 Species-Habitat Associations
Altogether 47 tree and shrub species, accounting for 99.7% of all individuals, were available
to identify possible species-habitat associations
within the research plot. Out of 188 potential species-habitat associations (four habitat types × 47
species), 114 were signiicantly positive or negative based on the habitat randomization procedure. The randomized-species model reveals
139 signiicant associations, whereas the PCM
only shows 37 associations which are signiicant.
The PCM process identiies comparatively few
associations, but is much more realistic than the
randomized-species and habitat randomization
processes which do not consider the spatial autocorrelation of species and habitat, respectively.
Based on the PCM model, 34 out of 47 species
were signiicantly associated with one or more
habitat types, only two of these were negative.
The habitat randomization produced 39, the species randomization all 47 species with signiicant
associations (Table 3; see also Appendix 6, 7
and 8). These associations were positive as well
as negative. Again, habitat randomization and
the randomized-species approach produced an
inlated number of signiicant associations. The
PCM model appears to give more realistic results.
Table 4 shows four cross tabulations with the
numbers of common species-habitat associations,
based on the PCM model. Few species are signiicantly associated with more than one habitat.
Most habitat pairs have zero common specieshabitat associations which are either signiicantly
positive or negative. Signiicant associations
across several habitat types are very rare.
Table 4. Cross tabulations of common species-habitat associations based on the PCM. The symbol “+” indicates
signiicant positive, “−” signiicant negative associations; N indicates no signiicant associations.
Habitat 1+
Habitat 1N
Habitat 1−
Habitat 1+
Habitat 1N
Habitat 1−
Habitat 1+
Habitat 1N
Habitat 1−
508
Habitat 2+
Habitat 2N
Habitat 2−
0
4
1
16
26
0
0
1
0
Habitat 3+
Habitat 3N
Habitat 3−
0
5
0
16
25
1
0
0
0
Habitat 4+
Habitat 4N
Habitat 4−
0
10
0
16
20
1
0
0
0
Habitat 3+
Habitat 2+
Habitat 2N
Habitat 2−
1
4
1
Habitat 4+
Habitat 2+
Habitat 2N
Habitat 2−
1
9
0
Habitat 4+
Habitat 3+
Habitat 3N
Habitat 3−
0
10
0
Habitat 3N
4
38
0
Habitat 4N
3
33
1
Habitat 4N
5
32
0
Habitat 3−
0
0
0
Habitat 4−
0
0
0
Habitat 4−
0
0
0
Zhang et al.
4 Discussion
Habitat specialization is a prominent topic used to
explain particular patterns of species coexistence
in a forest community. Several studies in tropical
forests could identify some species-habitat associations, but similar investigations seem to be
lacking for temperate forest ecosystems. Accordingly, the purpose of this study is to broaden
our understanding of the structuring forces in a
temperate multi-species forest ecosystem with
particular reference to species-habitat associations based on topographic features.
We found that spatial autocorrelation of species
and topographic variables may confuse the contribution of topographic variation to plant spatial
patterns. The spatial distributions of most of the
tree and shrub species in our experimental plot
were distinctly aggregated. Species autocorrelation may be the result of seed-dispersal limitations of most tree species. For this reason, it was
necessary to reduce the effects of autocorrelation,
regarding species as well as habitat.
4.1 Habitat Types and Indicator Species
The experimental plot was subdivided into four
habitat types sharing topographic characteristics.
Habitat type 1 has the largest area and occupies
the low altitudes. Almost half of all the signiicant species-habitat associations were identiied
in habitat type 1, when considering both habitat
and species autocorrelation.
The lower plateau is characterized by moist
and wet soil conditions. The cells on the lower
plateau and lower gentle slopes (habitat types
1 and 3) have two types of indicators: light and
moisture-demanding pioneer and shade tolerant understorey species, including Betula spp.,
Euonymus spp., Acer spp., Malus baccata (L.)
Borkh. and Syringa reticulata var. amurensis. The
high lying areas and steep slopes (habitat types
2 and 4) are characterized by well-drained soil
conditions. These habitat types have more climax
indicator species, such as Pinus koraiensis, Acer
mono, Tilia amurensis, and Quercus mongolicus
Fisch. Some harvesting has taken place in the
experimental area during the early 1960’s, but
details are not available.
Species-Habitat Associations in a Northern Temperate Forest in China
Based on the results obtained with the PCM
model, the percentage of species showing signiicant associations with habitat type 1 (45.9%)
and habitat type 4 (27.0%) is consistent with the
relative habitat areas in the two dominant habitat
types. Habitat type 1 occupies 47 per cent of the
total experimental area, habitat type 4 only 27
per cent. This almost exact match is likely to be
accidental. However, the bigger areas naturally
support a greater number of species which are
candidates for signiicant species-habitat associations. Habitat types 2 and 3 represent the transitional zones between the low and high plateau.
Again, the total percentage of habitat-associated
species in both types together (27.0%) closely
matches that of their total area (25.9%). Thus,
the number of habitat-associated species in a speciic habitat may depend on the available habitat
area. This is a species-area phenomenon. Due
to the increasing number of trees, more habitat
specialists can be found in habitat types that
occupy larger areas. This is simply an effect of
the species-area relationship. However, this effect
does not contradict the occurrence of species
specialization. The majority of the species in our
experimental area showed signiicant positive or
negative associations with speciic habitat types.
The complex topography in the study area
includes habitat types which are preferred by
different tree and shrub species, resulting in a high
proportion of distinct species-habitat associations.
Our indicator species analysis showed that 38 out
of 47 of the more common species were indicative
of speciic habitat types. The indicative power
was rather variable among the different indicator
species. For example, in habitat type 1, Ulmus
davidiana var. japonica showed the highest indicator value (0.62), while Betula platyphylla Suk.
had the lowest indicator value (0.10).
4.2 Species-Habitat Associations
Almost all the studies about species-habitat associations, based on data from large experimental
areas, were conducted in tropical or subtropical
forest ecosystems. Harms et al. (2001) identiied
six habitat types in the tropical forest in Barro
Colorado Island, where many species had strong
species-habitat associations. Similar results were
509
Silva Fennica 46(4), 2012
obtained in other tropical forests (Harms et al.
2001, Itoh et al. 2003, Russo et al. 2005, Yamada
et al. 2006, 2007, John et al. 2007). Our study
appears to be among the irst ones based on data
from a large experimental area in a temperate
forest ecosystem.
Altogether 34 species (72.3% of the total
number) are associated with at least one habitat
type, based on the conservative PCM model.
Interestingly, these 34 species only produce 35
signiicantly positive species-habitat associations
(out of 136 potential associations). Almost each
of these species shows a unique positive association with a particular habitat. Ulmus davidiana
var. japonica is an exception, being signiicantly
associated with two habitat types simultaneously.
The patterns of different species specializing in
the different habitat types can be interpreted by
the life history strategies (for example, shade tolerance, growth and mortality rates, etc). According to Nakashizuka et al. (1992) and Masaki et al.
(1999), shade-tolerant species have recruitment
rates that are almost equivalent to, or signiicantly larger than mortality rates. This means that
their populations can be maintained or may even
increase under natural conditions. Understory
shade-tolerant species, such as Lonicera ruprechtiana Regel., Rhamnus davurica Pall., Euonymus
macropterus Rupr., Viburnum sargenti Koehne,
are neutrally associated with all four habitat types
in the research plot. Some dominant canopy species, such as Tilia amurensis, Pinus koraiensis,
Acer mono and Ulmus spp. show a strong adaptability toward habitat variations. These species
did not indicate any particular habitat preference
and had neutral associations with all habitat types.
Species may differ from each other in their
habitat preference. They are presumed to be capable of locating themselves in different positions
along habitat gradients. For example, Fraxinus
mandshurica specializes on habitat type 1, while
Quercus mongolica clearly prefers habitat type
4. This corresponds to the ecological characteristics of the two species. Fraxinus mandshurica
prefers fertile, moist and well drained sites, and
usually occurs on the gentle slopes. The cells of
habitat type 1 are located on these sites. Quercus
mongolica prefers dry sites, and is usually found
on the hilltop and lower ridges. That is where the
cells of habitat type 4 are found.
510
research articles
The distribution of some species relects
regeneration following local disturbance. Betula
platyphylla, for example, is a light-demanding
and shade-intolerant tree species. The species is
positively associated with habitat type 1 where
heavy cutting disturbance occurred about 50 years
ago. Viburnum burejaeticum Regel et Herd which
is a distinctly light-demanding species, showed
a particular preference for habitat type 3 and a
negative association with habitat type 2. These
habitat types differ with regard to aspect. The cells
in habitat type 3 are located on sunny south-facing
slopes, while most cells in habitat type 2 are found
on rather more shaded south-westerly slopes.
Habitat differentiation and spatial limitations
due to species dispersal are two key factors that
contribute to structuring forest communities. The
habitat differentiation theory is based on the idea
that there is a trade-off between growth/survival
rates and resource availability (Kitajima 1994,
Wright et al. 2003), producing environmentally
dependent species preferences. This theory seems
to represent the existing mainstream view (Clark
et al. 1999, Svenning 2004, Harms et al. 2001,
Wright 2002). Several studies have shown that
forest community structure is governed by seed
dispersal limitation and demographic stochasticity (Hubbell et al. 1999, Bell 2001, Hubbell
2001). These indings are supported by evidence
of seed dispersal limitation (Hubbell et al. 1999,
Dalling et al. 2002) and nonenvironmental spatial
dependency in species distributions (Svenning
2001, Tuomisto et al. 2003).
The understanding of species-habitat associations depends largely on the ability to reduce the
effects of autocorrelation, regarding species as
well as habitat. Spatial autocorrelation cannot be
ignored when considering species-habitat associations (Legendre and Legendre 1998). Restricted
seed dispersal and autocorrelated habitat types
may create an artiicially inlated effect to specieshabitat associations. This study has shown that
randomizing species (under the CSR model) and
randomizing habitat types seriously overestimate
the number of distinct habitat associations. For
this reason, the more realistic and conservative
PCM model was also used. The PCM approach
identiied much less signiicant species-habitat
associations (only 37, see Table 3) than the species and habitat randomization methods (114 and
Zhang et al.
139 signiicant associations, respectively). Our
results indicate that habitat differentiation and
dispersal limitation are not mutually exclusive.
Both contributed simultaneously to the maintenance of the particular distribution of the tree and
shrub species in our experimental area.
A species which is positively associated with
a particular habitat can be expected to have a
greater competitive advantage than other species
that are neutrally or negatively associated with
the same habitat. According to an earlier result
presented by Harms et al. (2001) in a tropical
forest, if negative associations were used to identify disappearing (“sink”) subpopulations within
the research area, then the list of species neutrally
or positively associated with a particular habitat
type would be equal to the number capable of
sustaining populations if the plot was composed
of only that habitat type. According to our PCM
tests, 32 out of 47 species were neutrally or
positively associated with the four habitat types.
Habitat types 1 and 2 were negatively associated (avoided) by Acer tegmentosum Maxim. and
Viburnum burejaeticum respectively, based on
the PCM model. This leaves 46 species which
are neutrally or positively associated with either
habitat type 1 or 2. No negative species-habitat
associations were found in habitat type 3 and 4.
Habitat types are not closed systems, and there
are immigrations between them via seed dispersals on short- and/or long-distances. This suggests
that species assemblages will be inluenced signiicantly by species input from the surrounding
habitat types. Thus, the high proportion of species
in these assemblages might be maintained by persistent immigration, not by habitat differentiation
among coexisting species. Sink areas are subsidized by sources and even strong habitat preference of trees and shrubs does not seem to provide
suficient support for the hypothesis that habitat
differentiation maintains high species diversity.
Future studies of species-habitat associations
should therefore examine whether habitat, if not
subsidized by immigration, would be able to support non-negative tree population growth.
Species-Habitat Associations in a Northern Temperate Forest in China
Acknowledgements
This research is supported by the National Basic
Research Program of China (973 Program;
2011CB403205), the Fundamental Research
Funds for the Central Universities (HJ201019), the National Natural Science Foundation
of China (31200315), and the 12th ive-year
National Science and Technology plan of China
(2012BAC01B03).
References
Bell, G. 2001. Neutral macroecology. Science 293:
2413–2418.
Clark, D.B., Clark, D.A. & Read, J.M. 1998. Edaphic
variation and the mesoscale distribution of tree species in a neotropical rain forest. Journal of Ecology
86: 101–112.
— , Palmer, M.W. & Clark, D.A. 1999. Edaphic factors and the landscape-scale distributions of tropical rain forest trees. Ecology 80, 2662–2675.
Comita, L.S., Condit, R. & Hubbell, S.P. 2007. Developmental changes in habitat associations of tropical
trees. Journal of Ecology 95: 482–492.
Condit, R., Hubbell, S.P. & Foster, R.B. 1996. Changes
in tree species abundance in a neotropical forest:
impact of climate change. Journal of Tropical Ecology 12: 231–256.
Dalling, J.W., Muller-Landau H.C., Wright S.J. & Hubbell S.P. 2002. Role of dispersal in the recruitment
limitation of neotropical pioneer species. Journal
of Ecology 90: 714–727.
De’ath, G. 2002. Multivariate regression trees: a new
technique for modeling species–environment relationships. Ecology 83: 1105–1117.
— 2010. mvpart: multivariate partitioning. R package
version 1.3-1. Available at: hhttp://cran.r-project.
org/.
Diggle, P. 1983. Statistical analysis of spatial point
patterns. Academic, London.
Dufrêne, M. & Legendre P. 1997. Species assemblages and indicator species: the need for a lexible
asymmetrical approach. Ecological Monographs
67: 345–366.
Harms, K.E., Condit, R., Hubbell, S.P. & Foster, R.B.
2001. Habitat associations of trees and shrubs in a
50-ha neotropical forest plot. Journal of Ecology
511
Silva Fennica 46(4), 2012
89: 947–959.
Hubbell, S.P. 2001. The uniied neutral theory of biodiversity and biogeography. Monographs in population biology. Princeton University Press, Princeton,
New Jersey, USA.
— , Foster, R.B., O’Brien, S.T., Harms, K.E., Condit,
R., Wechsler, B., Wright, S.J. & Loo & de Lao, S.
1999. Light-gap disturbances, recruitment limitation, and tree diversity in a Neotropical forest.
Science 283: 554–557.
Itoh, A., Yamakura, T., Ohkubo, T., Kanzaki, M., Palmiotto, P.A., LaFrankie, J.V., Ashton, P.S. & Lee, H.S.
2003. Importance of topography and soil texture in
the spatial distribution of two sympatric dipterocarp
trees in a Bornean rainforest. Ecological Research
18: 307–320.
John, R., Dalling, J.W., Harms, K.E., Yavitt, J.B., Stallard, R.F., Mirabello, M., Hubbell, S.P., Valencia,
R., Navarrete, H., Vallejo, M. & Foster, R.B. 2007.
Soil nutrients inluence spatial distributions of tropical tree species. PNAS 104: 864–869.
Kitajima, K. 1994. Relative importance of photosynthetic traits and allocation patterns as correlates
of seedling shade tolerance of 13 tropical trees.
Oecologia 98: 419–428.
Lai, J.S., Mi, X.C., Ren, H.B. & Ma, K.P. 2009. Species-habitat associations change in a subtropical
forest of China. Journal of Vegetation Science 20:
415–423.
Legendre, P. & Legendre, L. 1998. Numerical ecology.
2nd English ed. Elsevier Science BV, Amsterdam.
Masaki, T., Tanaka, H., Tanouchi, H., Sakai, T. &
Nakashizuka, T. 1999. Structure, dynamics and
disturbance regime of temperate broad-leaved forests in Japan. Journal of Vegetation Science 10:
805–814.
Nakashizuka, T. 2001. Species coexistence in temperate, mixed deciduous forests. Trends in Ecology &
Evolution 16: 205–210.
Plotkin, J.B., Potts, M.D., Leslie, N., Manokaran, N.,
LaFrankie, J. & Ashton, P.S. 2000. Species-area
curves, spatial aggregation, and habitat specialization in tropical forests. Journal of Theoretical
Biology 207: 81–99.
Potts, M.D., Davies, S.J., Bossert, W.H., Tan, S. &
Supardi, M.N.N. 2004. Habitat heterogeneity and
niche structure of trees in two tropical rain forests.
Oecologia 139: 446–453.
R Development Core Team. 2011. R: a language and
environment for statistical computing. R Founda-
512
research articles
tion for Statistical Computing, Vienna, Austria.
ISBN 3-900051-07-0. Available at: http://www.Rproject.org/.
Roberts, D.W. 2010. labdsv: Ordination and multivariate analysis for ecology. R package version 1.4-1.
Available at: hhttp://cran.r-project.org/.
Russo, S.E., Davies, S.J., King, D.A. & Tan, S. 2005.
Soil-related performance variation and distributions of tree species in a Bornean rain forest.
Journal of Ecology 93: 879–889.
Svenning, J.-C., Kinner, D.A., Stallard, R.F., Engelbrecht, B.M.J. & Wright, S.J. 2004. Ecological
determinism in plant community structure across a
tropical forest landscape. Ecology 85: 2526–2538.
— 2001. Environmental heterogeneity, recruitment
limitation, and the mesoscale distribution of palms
in a tropical montane rain forest (Maquipucuna,
Ecuador). Journal of Tropical Ecology 17: 97–113.
Tuomisto, H., Ruokolainen, K. & Yli-Halla, M. 2003.
Dispersal, environment, and loristic variation of
western Amazonian forests. Science 299: 241–244.
Webb, C.O. & Peart, D.R. 2000. Habitat associations of
trees and seedlings in a Bornean rain forest. Journal
of Ecology 88: 464–478.
Wright, S.J. 2002. Plant diversity in tropical forests:
a review of mechanisms of species coexistence.
Oecologia 130: 1–14.
— , Muller-Landau, H.C., Condit, R. & Hubbell, S.P.
2003. Gap-dependent recruitment, realized vital
rates, and size distributions of tropical trees. Ecology 84: 3174–3185.
Yamada, T., Tomita, A., Itoh, A., Yamakura, T., Ohkubo,
T., Kanzaki, M., Tan, S. & Ashton, P.S. 2006. Habitat associations of Sterculiaceae trees in a Bornean
rain forest plot. Journal of Vegetation Science 17:
559–566.
— , Zuidema, P.A., Itoh, A., Yamakura, T., Ohkubo, T.,
Kanzaki, M., Tan, S. & Ashton P.S. 2007. Strong
habitat preference of a tropical rain forest tree does
not imply large differences in population dynamics
across habitats. Journal of Ecology 95: 332–342.
Yamakura, T., Kanzaki, M., Ohkubo, T., Ogino, K.,
Chai, E.O.K., Lee, H.S. & Ashton, P.S. 1995.
Topography of a large-scale research plot established within a tropical rain forest at Lambir,
Sarawak. Tropics 5: 41–56.
Zhang, C., Gao, L. & Zhao, X. 2009. Spatial structures in a secondary forest in Changbai Mountains
Northeast China. Allgemeine Forst- und Jagdzeitung (AFJZ) 180: 45–55.
Zhang et al.
Species-Habitat Associations in a Northern Temperate Forest in China
— , Zhao, X. & Gadow, K. v. 2010. Partitioning
temperate plant community structure at different
scales. Acta Oecologica 36: 306–313.
Total of 36 references
Appendix 1. The total basal area of dominant tree species in
research plot.
Tree species
Ulmus davidiana var. japonica
Pinus koraiensis
Juglans mandshurica
Tilia mandschurica
Carpinus cordata
Acer mono
Fraxinus mandshurica
Tilia amurense
Ulmus laciniata
Total basal area
(m2/ha)
0.4309
0.6833
1.8705
0.6098
0.3813
1.2182
0.6023
0.8771
0.9912
Appendix 2. Stem density of the top ive species in research plot.
Tree species
Acer mandshuricum
Syringa reticulata var. amurensis
Ulmus davidiana var. japonica
Carpinus cordata
Acer mono
Stem density
(stems/ha)
150
412
141
393
240
513
Silva Fennica 46(4), 2012
research articles
Appendix 3. List of the 47 species used in the analysis
Species*
Family
Life forms
Betula platyphylla
Acer mandshuricum
Syringa reticulata var. amurensis
Acer ginnala
Lonicera ruprechtiana
Rhamnus davurica
Euonymus macropterus
Padus racemosa
Ulmus davidiana var. japonica
Acanthopanax senticosus
Acer barbinerve
Ulmus macrocarpa
Berberis amurensis
Lonicera maackii
Philadelphus schrenkii
Rhamnus schneideri var. manshurica
Betula costata
Pinus koraiensis
Juglans mandshurica
Fraxinus rhynchophylla
Maackia amurensis
Phellodendron amurense
Viburnum sargenti
Tilia mandshurica
Ulmus laciniata
Euonymus verrucosus
Aralia elata
Corylus mandshurica
Quercus mongolica
Acer trilorum
Viburnum burejaeticum
Carpinus cordata
Acer tegmentosum
Actinidia arguta
Acer mono
Abies holophylla
Malus baccata
Pyrus ussuriensis
Populus davidiana
Fraxinus mandshurica
sorbus alnifolia
Euonymus alatus
Rhamnus ussuriensis
Rhamnus parvifolia
Rhamnus diamantiaca
Lonicera praelorens
Tilia amurensis
Betulaceae
Aceraceae
Oleaceae
Aceraceae
Caprifoliaceae
Rhamnaceae
Celastraceae
Rosaceae
Ulmaceae
Araliaceae
Aceraceae
Ulmaceae
Berberidaceae
Caprifoliaceae
Saxifragaceae
Rhamnaceae
Betulaceae
Pinaceae
Juglandaceae
Oleaceae
Leguminosae
Rutaceae
Caprifoliaceae
Tiliaceae
Ulmaceae
Celastraceae
Araliaceae
Betulaceae
Fagaceae
Aceraceae
Caprifoliaceae
Betulaceae
Aceraceae
Actinidiaceae
Aceraceae
Pinaceae
Rosaceae
Rosaceae
Salicaceae
Oleaceae
Rosaceae
Celastraceae
Rhamnaceae
Rhamnaceae
Rhamnaceae
Caprifoliaceae
Tiliaceae
Tree
Tree
Small tree/tree
Shrub/small tree
Shrub
Shrub/small tree
Shrub
Tree
Tree
Shrub
Small tree
Shrub/tree
Shrub
Shrub
Shrub
Shrub
Tree
Tree
Tree
Tree
Tree
Tree
Shrub
Tree
Tree
Shrub
Shrub
Shrub
Tree
Tree
Shrub
Tree
Tree
Liana
Tree
Tree
Tree
Tree
Tree
Tree
Tree
Shrub
Shrub
Shrub
Shrub
Shrub
Tree
* The species were identiied using the records in the Chinese Virtual Herbarium (see http://www.cvh.org.cn/cms/)
514
Zhang et al.
Species-Habitat Associations in a Northern Temperate Forest in China
Appendix 4. Multivariate regression tree for the species composition data. Euclidean distance was used for splitting.
Barplots show the multivariate species mean at each node, and the numbers of sites are shown in parentheses.
515
Silva Fennica 46(4), 2012
research articles
Appendix 5. Results of indicator species analysis.
516
Species
Habitat type
Indicator value
Ulmus davidiana var. japonica
Acer trilorum
Syringa reticulata var. mandshurica
Lonicera maackii
Malus baccata
Fraxinus mandshurica
Rhamnus davurica
Padus racemosa
Juglans mandshurica
Euonymus macropterus
Euonymus alatus
Phellodendron amurense
Acer ginnala
Maackia amurensis
Betula platyphylla
Carpinus cordata
Pinus koraiensis
Actinidia arguta
Acer tegmentosum
Aralia elata
Acer mandshuricum
Philadelphus schrenkii
Acer barbinerve
Abies holophylla
Rhamnus schneideri var. manshurica
Euonymus macropterus
Betula costata
Acer mono
Sorbus alnifolia
Ulmus laciniata
Tilia mandshurica
Tilia amurensis
Corylus mandshurica
Fraxinus rhynchophylla
Quercus mongolica
Rhamnus parvifolia
Lonicera praelorens
Acanthopanax senticosus
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 1
Habitat 2
Habitat 2
Habitat 2
Habitat 2
Habitat 2
Habitat 3
Habitat 3
Habitat 3
Habitat 3
Habitat 3
Habitat 3
Habitat 3
Habitat 4
Habitat 4
Habitat 4
Habitat 4
Habitat 4
Habitat 4
Habitat 4
Habitat 4
Habitat 4
Habitat 4
Habitat 4
0.6178
0.6054
0.5329
0.4643
0.4508
0.3961
0.3574
0.353
0.3134
0.2823
0.2693
0.2429
0.1157
0.1147
0.1007
0.5743
0.2395
0.2172
0.1769
0.0403
0.2847
0.2149
0.1892
0.1184
0.117
0.1117
0.0961
0.3386
0.3122
0.31
0.2752
0.2721
0.2457
0.2106
0.1983
0.0663
0.0641
0.0453
Probability
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.006
0.001
0.001
0.023
0.001
0.01
0.001
0.001
0.026
0.016
0.002
0.033
0.005
0.014
0.047
0.002
0.003
0.001
0.001
0.001
0.025
0.008
0.001
0.026
0.004
0.032
0.05
Zhang et al.
Species-Habitat Associations in a Northern Temperate Forest in China
Appendix 6. Species-habitat associations using the PCM model. “p” and “n” denote to “positive” and
“negative”, respectively. *, ** and *** indicate signiicance at p < 0.05, 0.01 and 0.001, respectively.
Species
Habitat 1
Betula platyphylla
Acer mandshuricum
Syringa reticulata var. amurensis
Acer ginnala
Lonicera ruprechtiana
Rhamnus davurica
Euonymus macropterus
Padus racemosa
Ulmus davidiana var. japonica
Acanthopanax senticosus
Acer barbinerve
Ulmus macrocarpa
Berberis amurensis
Lonicera maackii
Philadelphus schrenkii
Rhamnus schneideri var. manshurica
Betula costata
Pinus koraiensis
Juglans mandshurica
Fraxinus rhynchophylla
Maackia amurensis
Phellodendron amurense
Viburnum sargenti
Tilia mandshurica
Ulmus laciniata
Euonymus verrucosus
Aralia elata
Corylus mandshurica
Quercus mongolica
Acer trilorum
Viburnum burejaeticum
Carpinus cordata
Acer tegmentosum
Actinidia arguta
Acer mono
Abies holophylla
Malus baccata
Pyrus ussuriensis
Populus davidiana
Fraxinus mandshurica
Sorbus alnifolia
Euonymus alatus
Rhamnus ussuriensis
Rhamnus parvifolia
Rhamnus diamantiaca
Lonicera praelorens
Tilia amurensis
p*
Habitat 2
Habitat 3
Habitat 4
p***
p***
p***
p***
p**
p***
p***
p*
p*
p***
p***
p***
p***
p*
p**
p*
p*
p**
p*
p***
n**
n***
p***
p***
p***
p*
p*
p***
p*
p***
p**
p***
p*
p***
p***
p***
517
Silva Fennica 46(4), 2012
research articles
Appendix 7. Species-habitat associations using the CRS model. “p” and “n” denote to “positive” and “negative”, respectively. *, ** and *** indicate signiicance at p < 0.05, 0.01 and 0.001, respectively.
Species
Habitat 1
Habitat 2
Betula platyphylla
Acer mandshuricum
Syringa reticulata var. amurensis
Acer ginnala
Lonicera ruprechtiana
Rhamnus davurica
Euonymus macropterus
Padus racemosa
Ulmus davidiana var. japonica
Acanthopanax senticosus
Acer barbinerve
Ulmus macrocarpa
Berberis amurensis
Lonicera maackii
Philadelphus schrenkii
Rhamnus schneideri var. manshurica
Betula costata
Pinus koraiensis
Juglans mandshurica
Fraxinus rhynchophylla
Maackia amurensis
Phellodendron amurense
Viburnum sargenti
Tilia mandshurica
Ulmus laciniata
Euonymus verrucosus
Aralia elata
Corylus mandshurica
Quercus mongolica
Acer trilorum
Viburnum burejaeticum
Carpinus cordata
Acer tegmentosum
Actinidia arguta
Acer mono
Abies holophylla
Malus baccata
Pyrus ussuriensis
Populus davidiana
Fraxinus mandshurica
Sorbus alnifolia
Euonymus alatus
Rhamnus ussuriensis
Rhamnus parvifolia
Rhamnus diamantiaca
Lonicera praelorens
Tilia amurensis
p***
n***
p***
p***
n**
n***
n***
n***
n***
518
p***
p***
p***
p***
n*
n***
p*
n***
p***
p***
n***
n***
p***
n***
p***
p***
n***
n***
n***
p***
n***
n***
n***
p***
p***
n***
n***
n***
n***
p***
p**
n***
p***
n***
p***
n***
p***
n***
Habitat 3
p***
n***
n***
n***
n***
p***
n***
n*
p***
n***
n***
n***
p***
p***
n***
p*
n***
n***
p***
n*
n*
n***
p***
n***
p***
n***
n***
p***
p***
p***
p***
n*
n***
p***
p***
n***
n***
p*
p***
n***
n***
n***
n***
n***
p***
p***
n**
p***
p**
n***
p***
n***
n***
p***
p***
p***
n***
p*
p***
p***
n***
n*
p***
p***
p***
n***
n***
p***
n***
p***
n***
n***
n***
p***
n***
p***
n***
n***
n***
p**
p***
p***
p***
n***
n***
n***
Habitat 4
p***
n*
p***
n***
p***
p***
Zhang et al.
Species-Habitat Associations in a Northern Temperate Forest in China
Appendix 8. Species-habitat associations based on habitat randomization. “p” and “n” denote to “positive”
and “negative”, respectively. *, ** and *** indicate signiicance at p < 0.05, 0.01 and 0.001, respectively.
Species
Habitat 1
Habitat 2
Habitat 3
Betula platyphylla
Acer mandshuricum
Syringa reticulata var. amurensis
Acer ginnala
Lonicera ruprechtiana
Rhamnus davurica
Euonymus macropterus
Padus racemosa
Ulmus davidiana var. japonica
Acanthopanax senticosus
Acer barbinerve
Ulmus macrocarpa
Berberis amurensis
Lonicera maackii
Philadelphus schrenkii
Rhamnus schneideri var. manshurica
Betula costata
Pinus koraiensis
Juglans mandshurica
Fraxinus rhynchophylla
Maackia amurensis
Phellodendron amurense
Viburnum sargenti
Tilia mandshurica
Ulmus laciniata
Euonymus verrucosus
Aralia elata
Corylus mandshurica
Quercus mongolica
Acer trilorum
Viburnum burejaeticum
Carpinus cordata
Acer tegmentosum
Actinidia arguta
Acer mono
Abies holophylla
Malus baccata
Pyrus ussuriensis
Populus davidiana
Fraxinus mandshurica
Sorbus alnifolia
Euonymus alatus
Rhamnus ussuriensis
Rhamnus parvifolia
Rhamnus diamantiaca
Lonicera praelorens
Tilia amurensis
p***
n**
n***
n***
n**
n*
n**
n***
n***
n**
n***
n***
n**
n***
p***
p***
p***
p***
n***
n***
n***
n***
n***
n***
p***
p***
n***
n***
p***
n**
n***
n***
p***
p*
n***
n***
n***
n***
n***
n***
n***
n***
n***
p***
p***
p*
n***
p*
p***
n***
n***
n**
n***
n***
p***
p***
p***
n**
n***
n***
p***
p***
n***
n***
n***
p***
p**
p***
p*
n***
n*
p***
p***
n**
Habitat 4
n***
n***
n***
p**
p***
p***
n***
n*
n**
n***
p***
p**
p***
p***
p***
n***
n***
n***
n***
p***
n***
p***
n***
n***
n***
n***
n***
n***
p***
n**
p***
n***
n***
n***
n***
p***
n**
p***
p***
p***
519