|
Comparative analysis of primary subtropical rain
forests under different climatic conditions 異なる気候環境下に成立した原生的亜熱帯林の植生比較 －西表島と沖縄本島における植生調査－ |
|
||||||||||||||||||||||
Contents
English version Ⅰ．Summary Ⅱ．Introduction Ⅲ．Study area Ⅳ．Methods Ⅴ．Analysis method Ⅵ．Results Ⅶ．Discussion Japanese version Ⅷ．はじめに Ⅸ．調査地 Ⅹ．調査方法 ⅩⅠ．解析方法 ⅩⅡ．結果 ⅩⅢ．考察 ⅩⅣ．摘要 その他
ⅩⅤ．Acknowledgement ⅩⅥ．Literature cited |
English version
Ⅰ. Summary The effect of
different climatic conditions on the structure of community was clarified by
investigating primary sub-tropical rain forest in Okinawa Island and Iriomote
Island, southern Japan. In this study the results from species composition
show a higher floristic richness in the Ryukyu Islands, comparing with
Honnsyu. In the former region some
sub-tropical species grew of 52 – 61 %. Moreover, the species on Mt. Komi in Iriomote Island contained
the species of Yaeyama Island. The results suggested that the vegetation of the
Okinawa and Iriomote plots had not only typical species of Honnsyu but
also many sub-tropical characteristic species. Due to our findings, the
vegetation of our study areas in Ryukyu Island should be considered as pure
sub-tropical rain forest. And the species diversity of those areas was as
high as that of tropical rain forests in Thailand and Panama (1/D =
26.5, 1-D = 0.96 in Mt. Yonaha plot, 1/D =22.3, 1-D = 0.96
in Mt. Komi plot). Within the Ryukyu Islands the two plots, Mt. Yonaha in
Okinawa Island and Mt. Komi in Iriomote Island, had almost similar
vegetation. The remarkable difference between two plots was the relative
frequency of each dominant species. In Mt. Yonaha, Castanopsis sieboldii occupied
with high relative biomass, 52 %. In contrast, Castanopsis sieboldii assumed
for the same ratio, 26 % as Quercus miyagii in Mt. Komi. By comparing each plot, the diversity of Mt. Komi was higher than
that of Mt. Yonaha especially in upper layers. The size structure of the each
dominant species explained that the canopy species of Mt. Komi grew up
bigger and higher than that of Mt. Yonaha. If canopy species grew up bigger and higher, the space under the
canopy trees expanded and sub-canopy species could spread branches and leaves. Moreover sub-canopy species, which get into
higher layer, increase, and thus the diversity of Mt. Komi become higher in
upper layer. Complex spatial structure among the species divide three layers such
as canopy species, sub-canopy species and understory species and two patterns
like the species with aggregated distribution and species with random
distribution. The correlation coefficients of
spatial distribution between the component species explained that the more
the pattern of spatial distribution of species differ, the more component
species coexist. It was concluded that there is relationship such as
co-existence and competition of species behind the spatial structure, and
their relationship produce the high species diversity in Mt. Yonaha plot and
Mt. Komi plot. Key Words : Sub-tropical
forest, Species diversity, Castanopsis
sieboldii, Quercus miyagii, Size
structure, Spatial structure. Ⅱ. Introduction A vegetation type is directly affected by
the regional climatic condition. It is generally accepted that vegetation
pattern adapts to the change of climate along latitude. As Japan runs from
north latitude 46 to south latitude 24, the vegetation zones are divided into
three areas such as evergreen broad-leaved forests, deciduous broad-leaved
forests and boreal coniferous forests. Especially in southwestern Japan, most
of climax laurel forests consist of common dominant taxa like Machilus,
Quercus and Castanopsis (Tadaki 1971). However, the regional
vegetation is influenced by not only climatic conditions but also topography
and biological conditions, resulting in the landscape of many different
forest communities. For example, despite of most of the vegetation in
southwestern Japan are defined as Camellietea japonicae, the
vegetation from Yaku Island to northern Okinawa Island has been classified as
Psychotria rubra- Castanopsis sieboldii community (Miyawaki, 1980). Moreover, the vegetation of Iriomote
Island is defined as Quercus miyagii-Symplocos cochinchinensis community
(Hattori, Miyawaki, 1987). Those vegetation units
according to the plant sociological classification are different in species
composition, growth condition and physiognomy (Numata, 1974). It is important
to elucidate the maintenance mechanisms of species diversity at community
level for the conservation of diversity in forests (Kimoto & Takeda,
1989). Then, explaining the pattern of community and species diversity from
the number of tree and the relative frequency of component species in
community is important for community ecology. Ohyama and Yamamori (1971)
pointed out that complex topography play an important role in the high
floristic diversity in Yannbaru area, Okinawa prefecture. Miyawaki, Tagawa
and Kawakubo (1989) show the difference of species diversity, which is due to
the slight difference between the environment conditions of mountainous and
coastal regions in Iriomote Island. Moreover Suzuki, Kimoto and Takeda (1987)
explained that the component of salt resistance species reflected those of
geographic distribution by investigating the relationship between species
distribution and salt resistant of their species. Those researchers on
community ecology, were interested in three problems as follows (1) interspecific
competition, (2) population dynamics, (3) the classification of forest
community (Kimoto & Takeda, 1989). Although it was difficult to show them
clearly, this paper was trying to understand them by investigating primary
forests of Islands. This paper selected Okinawa Island and Iriomote Island in the Ryukyu Islands for study areas. These areas are located in transition zone from warm-temperate to tropical zone, and the forests consist of sub-tropical rain forests. These sub-tropical rain forests are limited to Ryukyu and Taiwan because other same latitude areas are arid zones such as desert, grassland and rain-green climate. However, primary forests are limited to northern part of Okinawa Island and Iriomote Island, because of raised coral reef effect and environment destruction (Nomura & Satou, 1963). The purpose of this study is to examine the effect of different climatic conditions on the structure of community by investigating primary forests. The point of view in this study is as follows: (1) What is the main difference between vegetations under different climatic conditions? (2) How do the different vegetations affect the diversity? (3) What kind of spatial structure is the forest assumed? This paper investigated the differences between both areas by tree census and comparing species composition, species diversity, size structure and spatial structure. In addition to that, I examine briefly the floristic relationship between the two Ryukyu forest and those laurel forest growing in Hosyuu area. Ⅲ. Study Area
The study was conducted in Mt. Yonaha,
northern Okinawa main Island and Mt. Komi, eastern Iriomote Island. Mt. Yonaha is 498 m above the sea level and it is
the highest site of Okinawa main Island. The area near the summit is
designated as nature reserve. Many endangered species such as Sapheopipo
noguchii, Rallus okinawae and Cheirotonus jambar inhabit
this mountainous region and thus this area is regarded as a very important
nature conservation area (Itou, 1998). This climax forest is evergreen
broad-leaved forest and dominated by Castanopsis sieboldii, Distylium
racemosum, Schefflera octophylla and so on. The mean annual temperature of Okinawa main
Island is ２２．４ ℃, and the mean annual rainfall is 2036 ｍｍ. Then Mt. Komi is 470
m and it is the highest site in Iriomote Island. Iriomote Island is the
second largest Island in Okinawa prefecture and about 90% of the Island is
covered with primary forests. This area was designated for Iriomote National
Park in 1972. Many animals and plants species such as Felis iriomotensis,
Spilornis chella perplexas, Cuora flaromarginata evelynae Eumeces
kishinouyei, Heritiera littoralis and Satakentia liukiuensis were
reserved as the protected animals and plants. This climax forest is evergreen
broad-leaved forest. And Castanopsis sieboldii, Quercuｓ miyagii and
Perser thunbergii dominate in
the mountainous area. It is often showed that some clambering plants like Freycinetia
formosana climbed the stem. The mean annual temperature of Iriomote Island
is ２３.３ ℃, and the mean annual rainfall is 2397 ｍｍ. Ⅳ. Methods
A quadrat (66×40 m) was set up near the top
of Mt. Yonaha and divided into 165 small sub-quadrats of 4×4 ｍ in May 2000. A
quadrat (50×70 m) was set up near the top of Mt. Komi and divided into 35
small sub-quadrats of 10×１０ ｍ in September 2000.Trees more than 2 m in height
were identified and measured for DBH ( stem diameter at breast height ), in
each sub-quadrat. Ⅴ. Analysis Method
Simpson’s Diversity index
Diversity of community was
examined by Simpson’s Diversity index. This equation is dependent on
consistency probability of two cases when they are took out. Simpson’s index
of concentration, D was examined by
where ni and N were the number of species
and total number of tree of i-th species. Because Simpson’s
Diversity index, 1/D is sensitive to change the number of dominant
species, a reciprocal of 1/D is suitable for the diversity differences
in the inside of concentration. Furthermore Simpson’s Diversity, 1-D
is not affected by tree density, 1-D is suitable for the diversity
differences in the relation of concentrations. If the value of both 1/D
and 1-D is large, the diversity is high. Multiple linear regression analysis The factor of the
diversity change in each vertical layer is examined by multiple linear
regression analysis. This analysis defines the relationship between one
variable as the dependent variables, and other variables as the explanatory
variable. This analysis was examined by Ｙ＝Ａ１X１＋Ａ２Ｘ２＋…＋ＡｐＸｐ＋Ａ０ where Y was the dependent variables, meaning result and Ｘ１，Ｘ２，…，Ｘｐ
were the explanatory variable, meaning some factors and Ａ１，Ａ２，…，Ａｐ
were coefficient of regression, meaning relationship between the result and
the factors. If A is positive value, X will affect Y
positively. If A is negative value, X will affect Y
negatively. In this study, multiple linear regression analysis was conducted
for the diversity index, 1/D, in the stands vertical layer as the dependent
variables and the DBH increment, number of tree, number of species and
frequency of dominant species of each plots as the explanatory variable. Ｋｏｌｍｏｇｏｒｏｖ－Ｓｍｉｒｎｏｖ test The difference in the size
structure of Castanopsis sieboldii stands two plots was examined by
Ｋｏｌｍｏｇｏｒｏｖ－Ｓｍｉｒｎｏｖ test, i.e. K-S test. Cluster Analysis The Cluster analysis was
used to study the component species based on species-specific parameters.
This method arranges the same quality items in different quality types,
therefore it can sort some groups. If the combination distance among the
species is shorter, the items are similar in some characteristics. In this
study, 20 species, from two plots, were classified based on frequency of
trees, rate of coppice, maximum DBH, skewness of DBH, coefficient of variance
(CV) of DBH and population biomass. Morishita’s Iδ-index
where q was the number of quadrats and xi
was the number of species in each quadrat and T was the total number
of species. If the value is > 1, the population will be aggregated
distribution. If the value is < 1, the population will be random
distribution. Correlation coefficients
where X and Y were the mean of x and y
and Sx and Sy were standard deviation of x
and y. Correlation coefficients, r, always shows :－１≦ｒ≦１. r≒－１ means
strong negative correlation, r＜0 means weak
negative correlation, r≒0 means no
correlation, r＞0 means weak
positive correlation, r≒１ means
strong positive correlation. In this study, the spatial distribution of
20 species in the two plots was examined by correlation coefficients. Ⅵ. Result
（１） Species Composition of the forests In total 9725 individuals (/ha) belonging
to 265 species (/ha) were recorded in the plot of Mt. Yonaha in Okinawa
Island, and in total 10830 individuals (/ha) belonging 220 species (/ha) were
recorded in the plot of Mt. Komi in Iriomote Island. Forty-five species
appeared in both plots. The relative frequency of the total number of trees
in Mt. Yonaha was Castanopsis sieboldii 10%, Distylium racemosum,
Psychotria rubra and Camellia lutchuensis 6%, Ardsia quinquegona
5%, Ilex ficoidea 4% and other 63 species 3-0.04%. And that of Mt.
Komi was Ardsia quinquegona and Psychotria rubra 10%, Ardisia
sieboldii and Distylium racemosum 7%, Quercus miyagii 6%, Randia
canthioides and Castanopsis sieboldii 5% and other 70 species
3-0.03%. According to the number of trees and biomass, the dominant species
of Mt. Yonaha were Castanopsis sieboldii and Distylium racemosum, and
those of Mt. Komi were Castanopsis sieboldii and Quercus miyagii.
Castanopsis sieboldii of Mt. Yonaha was 83.50 cm in the maximum DBH and Persea
thunbergii of Mt. Komi was 77.48 cm in the maximum DBH (Table.1-a, -b and Table.2). （２） Diversity of the plots According to Simpson’s diversity
index, in the plot of Mt. Yonaha in Okinawa Island, the values of 1/D
and 1-D were 26.5 and 0.96 respectively, and in Mt. Komi plot,
Iriomote Island, that of 1/D and 1-D were 22.3 and 0.96
respectively (Fig.3). According to diversity index of
each vertical layer in the stands (Fig.4-a, -b), the diversity index of Mt.
Yonaha was highest value (1/D = 40) at DBH = 10 cm and that of Mt. Komi was
not fluctuating and steadily. It is useful to use
multiple linear regression analysis in order to detect the factor of the
change of species diversity. As a result, the change of diversity of Mt.
Yonaha was significantly affected by the degree of the number of Castanopsis
sieboldii (P<0.05). And that of Mt. Komi was not substantially
affected. Comparing between plots,
the species diversity of Mt. Komi is higher than that of Mt. Yonaha
especially in upper layers (DBH > 15 cm). （３） Size structure of dominant species According to the order in terms of DBH of
individual trees, Castanopsis sieboldii in each plot could showed a
boundary between 10 cm and 15 cm in DBH because of the high rate of coppice,
78 % in Mt. Yonaha and 80 % in Mt. Komi (Fig.5-a –b –c –d). The boundary
among vertical layers in Mt. Yonaha, Okinawa Island was about 10 cm in DBH,
and that of Mt. Komi in Iriomote Island was near 15 cm in DBH. The values of
Mt. Komi were bigger than those of Mt. Yonaha, but not only the boundary but
also the highest layer of Castanopsis sieboldii (K-S test P>0.01).
Distylium racemosum of Mt. Yonaha reached the highest layer (about 35 cm
in DBH) while the layer of Quercus miyagii in Mt. Komi increased like
an exponential function. (4) Histogram Size class distributions of some
representative species in the two plots with typical distribution patterns
are presented in Fig.1-a, -b. These patterns were classified into three
groups as follows; (1) Understory species (DBH < 10 cm): Syzygium
buxifolium, Psychotria rubra, Randia canthioides, Ardsia
quinquegona and Antidesma japonicum (2) Sub-canopy species (DBH
< 30 cm): Ilex ficoidea, Schefflera octophylla, Ardisia sieboldii,
Elaeocarpus japonicus, Styrax japonicum, Persea japonica and Distylium
racemosum of Mt. Komi (3) Canopy species (appeared DBH > 30 cm): Castanopsis
sieboldii Distylium racemosum of Mt. Yonaha and Quercus miyagii
of Mt. Komi. Moreover, canopy species and sub-canopy species were
classified into two groups as follows; (1) L-shaped size class distribution: Persea
japonica, Distylium racemosum, Ilex ficoidea, Styrax japonicum and
Castanopsis sieboldii of Mt. Yonaha; (2) Scattered shaped size class
distribution: Schefflera octophylla, Ardisia sieboldii, Elaeocarpus
japonicus, and Castanopsis sieboldii of Mt. Komi. These patterns
of size class distribution except for Castanopsis sieboldii and Distylium
racemosum were similar in shapes. According to the classification of the
component species based on frequency of tree, rate of coppice, maximum DBH,
skewness of DBH, CV of DBH and population biomass, understory species of Mt.
Komi were significantly classified to different layers’ species, but a border
between understory and sub-canopy species of Mt. Yonaha was not clear. Each
canopy species was significantly classified (Fig.2-a –b). (5) Iδ-index of some representative species It is useful to detect the change of the
value of Iδ in order to investigate horizontal distribution of the
species (Fig.6-a –b –c). The horizontal distribution of species were
classified into two groups as follows; (1) the species with aggregated distribution:
Styrax japonicum, Persea japonica, Persea thunbergii, Antidesma
japonicum, Elaeocarpus japonicus, Viburnum japonicum and Castanopsis
sieboldii (Fig.6-a); (2) the species with random distribution: Psychotria
rubra, Distylium racemosum, Ilex ficoidea, Randia canthioides, Camellia
lutchuensis, Ilex maximowicziana and Euonymus japonicus (Fig6-b). (6) Correlation coefficients of spatial distribution between the
component species. Correlation coefficients of spatial
distribution between the component species belonging to same layer or same
horizontal distribution in the plots showed negative tendency. Such tendency
was recognized for the species of same layer, such as Randia
canthioides－Antidesma japonicum of Mt.Yonaha, Ilex ficoidea－Styrax
japonicum and Ardsia quinquegona－Antidesma japonicum of Mt. Komi
(Table.4). Ⅶ. Discussion （１） Definition of vegetation of plots The plant sociological studies could be
confused with respect to the classification of laurel forest from Ryukyu and
Honnsyu areas in southern Japan. In one side, this vegetation type has been
recognized between sub-tropical rain forests and warm-temperate rain forests.
But, in the other side, some authors consider both forests as the same
community (Yoshira, 1989). In this study the results from species composition
show a higher floristic richness in the Ryukyu Islands, comparing with
Honnsyu. In the former region some
sub-tropical taxa grew as Syzygium
buxifolium, Ilex ficoidea, Randia canthioides, Tricarysia dubia, Ilex
goshiensis, Antidesma japonicum, Euonymus lutchuensis. Moreover, the species on Mt. Komi in Iriomote Island contained
the species of Yaeyama Island: Pithecellobium lucidum, Hydrangea
yayeyamensis, Eurya yaeyamensis and Illicium tashiroi (Fig.2 –a, -b).
The results suggested
that the vegetation of the Okinawa and Iriomote plots had not only typical
species of Honnsyu: Castanopsis sieboldii, Distylium racemosum, Perser
thunbergii, Ligustrum japonicum, Persea japonica but also many sub-tropical
characteristic species. Due to our findings, the vegetation of our study
areas in Ryukyu Island should be considered as pure sub-tropical rain forest.
Within the Ryukyu Islands the two plots,
Mt. Yonaha in Okinawa Island and Mt. Komi in Iriomote Island, had almost
similar vegetation. The remarkable difference between two plots was the
relative frequency of each dominant species. In Mt. Yonaha, Castanopsis
sieboldii occupied with high relative biomass, 52 %. In contrast, Castanopsis
sieboldii assumed for the same ratio, 26 % as Quercus miyagii in
Mt. Komi. It was considered that the different dominant species could be
caused by the different soil and moisture conditions. Hattori (1985)
indicated the different pattern in distribution of species arise from
species-specific sea breeze salt tolerance. As Suzuki (1987) has stressed,
the lower sea breeze effects from coastal to inland, the higher species
diversity become. Some representative species, Syzygium buxifolium,
Symplocos prunifolia, Myrsine seguinii, Daphniphyllum teijsmannii, Myrica
rubra, Cinnamomum doederleinii, Ilex integra, Ternstroemis gymananthera, and
Camellia japonica, which have salt resistance. Symplocos
cochinchinensis, Wendlandia formosana, Distylium racemosum, Quercus miyagii,
Lasianthus fordii, Litsea acuminata, Tutcheria virgata, Ardisia sieboldii,
Lasianthus obliquinervis, Pithecellobium lucidum and Illicium tashiroi
seemed to be without sea breeze tolerant. Compared to these stands on the
basis of above-mentioned, in Mt. Yonaha, mountainous and coastal species
occupied 34 % and 6 % respectively. In Mt. Komi, mountainous and coastal
species occupied 36 % and 4 % respectively. Though both plots were located at
inland area near the summit, it was considered that the soil and moisture
conditions in Mt. Komi were better than in Mt. Yonaha. It seems that
mountainous species such as Quercus miyagii occur profitably at
suitable condition like a Mt. Komi plot, and thus some coastal species such
as Castanopsis sieboldii occur unfavorably. As a result the above
different dominant between the two plots may be caused. （２） Species diversity
The species
diversity index is used for evaluating richness of plots and comparing with
other forests. Ito (1998) used this index and pointed out the high diversity
of Yannbaru area, Okinawa pref. And he mentioned that many endangered animals
are able to survive in the high diversity forests. Noma (1997) mentioned that
the more kinds of fruit the forests have, the more kinds of animal are able
to live based on the relationship between animal-dispersed plants and
frugivores. These studies suggested the importance of high species diversity.
Due to our findings, the diversity index of plots in Mt. Yonaha, Okinawa
Island and Mt. Komi, Iriomote Island are as high as that of tropical forests
in Thailand and Panama (Fig.4). It
is important to consider how to coexist among the species in climax forest
when we explain the maintenance mechanisms of species diversity. It is said
that climax forest reaches a steady state consisting of shade tolerant trees
after long plant succession. However, as sunlight condition often changes in
the real environment, intolerant trees and annual-perennial herb are able to
coexist in climax forest. The change of sunlight condition is caused by the
various scale destruction due to a fallen of old tree and an effect of
typhoon. The area, which is improved sunlight condition, is called a gap. The
gap appears various scale and frequency, and produces the coexistence of
several species. Our study areas were often affected by typhoon on May and
June (Fig.5) and easy to cause some gaps. Therefore
it was guessed that the moderate effect of typhoon was a factor of high
diversity in this area. According
to diversity index of each vertical layer in the plots, the species diversity
of Mt. Yonaha was affected negatively by the density of Castanopsis
sieboldii. In other words, other species occupied equally except for Castanopsis
sieboldii. Whereas that of Mt. Komi was not substantially affected. It
was thought that these results were affected by the different relative
frequency of each dominant species. By
comparing each plot, the diversity of Mt. Komi was higher than that of Mt.
Yonaha especially in upper layers (Fig.7).
In order to confirm the result, the size structure of the each dominant
species was examined (Fig.8). In comparison with
common dominant species, Castanopsis sieboldii, the boundary among
vertical layers and the highest layer in Mt. Komi were higher than those of
Mt. Yonaha. Although it is inappropriate to regard the DBH of tree as growth
of tree in height, DBH are partly proportional relationship with tree height.
It seems that Castanopsis sieboldii of Mt. Komi was not limited to
grow more than that of Mt. Yonaha. Since other dominant species, Distylium
racemosum of Mt. Yonaha and Quercus miyagii of Mt. Komi, has
showed similar results, Quercus miyagii of Mt. Komi grew bigger and
higher. If such canopy species as Quercus miyagii and Castanopsis
sieboldii of Mt. Komi grew up bigger and higher, the space under the
canopy trees expanded and sub-canopy species could spread branches and leaves. Moreover sub-canopy species, which get into
higher layer, increase, and thus the diversity of Mt. Komi become higher in
upper layer. （３）Size structure and Spatial structure
Climax forests are composed of the species belonging to different spatial stages. There are some opinions about how to caused such complex stages. In one side, those species depend on gap and grow randomly in different seral stages (Watt, 1947). However, in the other side, most species regenerate under the closed forest without gap and grow with different maximum height (Aiba, 1994). Size class distribution of some
representative species in Mt. Yonaha, Okinawa Island and Mt. Komi, Iriomote
Island showed distinction between L-shape and scattered shaped (Fig.9). Considering the above mentioned, the L-shape
species are regenerating consecutively without gap and the scattered shaped
species begin to regenerate just after the light condition improves in gap.
Both of species co-occurred in this area. It seems that each species has own
regeneration strategy under different environment because the size class
distribution was not different between two plots. In the stand, the size
structure has extremely complicated and the boundary of each layer is not
clear. Therefore it is difficult to determine the position of layer from
height or DBH of tree (Aiba, 1994). In this study, the author tried to class
the layers with cluster analysis, whose variables based on maximum DBH, skewness
of DBH, CV of DBH, population biomass and so on (Fig.10).
The results suggest that there are some tendencies for the component species
to divide three layers such as canopy species, sub-canopy species and
understory species. Iδ-index explained the pattern of
spatial distribution (Fig.11). The results showed
that the component species were divided into two patterns like the species
with aggregated distribution and species with random distribution. The former
species regenerated with high rate of coppice (Fig.12).
But It was thought that some low rate of coppice species in the former
species such as Antidesma
japonicum are limited the range by
geographic range, soil moisture condition and so on. And it was considered
the latter species regenerated by seedling establishment without coppice.
These difference patterns suggest the different regeneration strategy among
species. The advantage of the coppice regeneration species is to be quick to
grow and cover the gap (Shinzato et al, 1986). In other side, seedling
regeneration species is superior to the variety of gene (Washitani, 1999). Fig.13 shows pattern of spatial distribution of the
component species. The correlation coefficients of spatial distribution
between the component species explained the importance of such complex
spatial structure (Table.4). The results suggest that
the co-occurring species in same layer are few and segregate their horizontal
distribution. In other words, the more the pattern of spatial distribution of
species differ, the more component species coexist. It was concluded that
there is relationship such as co-existence and competition of species behind
the spatial structure, and their relationship produce the high species
diversity in Mt. Yonaha plot and Mt. Komi plot. Japanese version
Ⅷ． はじめに 植生帯は、気候環境が変化すれば直接的な影響を受ける｡よって北から南に見られる植生配置のパターンは緯度傾度での気候が変動することに対応して変化してきた。南北に長い日本の場合､本州の中央部から南は常緑広葉樹林､それより北で北海道中央部までは落葉広葉樹林で､北海道の北東部は北方針葉樹林の植生帯にわかれる。特に､西南部では、どのような土地から遷移が発達しても、ほぼタブ･カシ･シイ類の優占する照葉樹林の極相に達するといわれている(只木,1971)。ただし、気候的条件が大前提としてあるとしても､その他に､土地条件､地形的条件､生物的条件などによって出来上がる極相には､多くの種類がある｡例えば､日本の照葉樹林は植生的にはヤブツバキクラス域として分類され、ヤブツバキ(Camellia
japonica)､ヒサカキ(Eurya japonica)､ヤブニッケイ(Cinnamomum doederleinii)などを標徴種とする植生が日本西南部の大部分を占めている。しかし、その中の屋久島から沖縄本島北部のヤンバル地域までの森林群集は､ボチョウジ－イタジイ群団と分類され（宮脇,1980）、その相観を形成する構成樹種はサクラツツジ(Rhododendron
tashiroi)､アデク(Syzygium buxifolium)､フカノキ(Schefflera octophylla),ボチョウジ(Psychotria
rubra)、シシアクチ(Ardisia quinquegona)､ギョクシンカ(Tarenna gracilipes)、オオシイバモチ(Ilex
ficoidea)などとされる。またさらに､西表島の潮風の影響を受けない適湿地ではオキナワウラジロガシ－アオバノキ群落が発達し(服部･宮城,1987)、オキナワウラジロガシ(Quercus
miyagii)の優占以外に､アオバノキ(Symplocos cochinchinensis)､オオシイバモチ､ホソバタブ(Persea
japonica)､バリバリノキ(Litsea acuminata)などが存在している。 以上に挙げたような群集は､特定の種組成･生育条件および相観をもち、植物社会学的群落分類における基本単位を意味しており(沼田,1974)､森林における生物多様性を保全するためには､この群集の種多様性がどのように維持されているのか､そのダイナミズムを解明する研究が必要不可欠といわれる(木元・武田,1989)。群集生態学では群集を構成する種数や構成種の相対的な個体数から、群集の規則性や多様性を説明することが重要課題となっている｡これまでの研究では、大山・山盛（1971）が、沖縄ヤンバル地方の亜熱帯林の種多様性には地形の複雑性が大きな役割を果たしていると指摘している。また、宮城､田川､川窪（１９８９）は、沖縄県西表島の南西端崎山半島の植生を対象とし､西表島主体部の植生との比較を行うことによって､主体部とは異なる植生であることを示した｡そして、島内においても東西南北で植生が異なることなどから、わずかな環境変化による群集の多様性も重要であることを指摘している｡同じく西表島において､鈴木、服部､武田（１９８７）は植生の分布と潮風の影響との関係について調べ､内陸へ入って潮風の影響が軽減するにともない種多様性は連続的に高くなっていくことを示した。そして、照葉樹林構成種は耐潮性の程度に差があり､いくつかの種群にまとめられること､構成種の地理的分布には耐潮性が反映していることを指摘している｡このような群集研究の興味ある対象としているのは､1)種間関係､2)複数の個体群の動態様式､3)特有な種の構成をもつ群集の境界、といった特徴である(木元・武田,1989)｡しかし､いかに明らかにするかはかなり困難で､本調査では､原生林に近い林分を調査の対象とすることで個体群の動態は競争－平衡状態が成立していると仮定し､境界については境界の明瞭な島を研究対象とすることで問題を単純化している｡ 本論の調査地の対象となったのは､南西諸島中部の沖縄島と､南端の西表島である。この地域は暖温帯から熱帯への移行帯に位置しており、そこには亜熱帯植生が成立している｡亜熱帯気候は琉球､台湾を除くとすべて（北メキシコ､南ブラジル､アラブ諸国､サハラ､南アフリカなど）、砂漠､草原､雨緑林などの乾燥地となるため､亜熱帯植生はこの地域に限定される｡また､これらの諸島を概観すると､乾燥しやすい隆起サンゴ礁などの土地条件や開発による破壊といった人工的な条件によって､原生林のままの形で森林が存在するところは少なく､沖縄本島北部と西表島にのみ見られるといわれている(野村･佐藤,1963)。 本論では､これまで報告されている植生調査資料を参考に、以上に挙げたような世界的にも大変貴重な地域である沖縄島と西表島の植生を調査し､気候環境の違いに応じて林木群集構造がどのように異なっているのかを明らかにする｡研究視点は､｢気候環境の違いが、どのように植生に影響するのか｣､「植生の違いにより、種多様性がどのように異なるのか」、｢森林はどのような空間構造で形成されているのか｣、｢その空間構造が意味することは何か｣などである｡具体的には､西表島･沖縄島のそれぞれの原生的林分にプロットを設定して毎木調査を行う｡そして､それぞれの種組成･種多様性･サイズ構造･空間構造などを比較し、これらの結果から、西表島･沖縄島それぞれの林木群集の構造を比較検討する｡ Ⅸ．調査地 本研究の調査区は沖縄本島北部の与那覇岳と､西表島東部の古見岳に設定した｡与那覇岳は標高498ｍと沖縄本島でもっとも高く､その頂上付近は天然記念物として保護されている｡山岳地帯には､ノグチゲラ(Sapheopipo
noguchii)、ヤンバルクイナ(Rallus okinawae)、ヤンバルテナガコガネ(Cheirotonus
jambar)をはじめ、多数の特殊な､絶滅に瀕した固有種がすんでおり､ここは自然保護の観点から世界的に見てもきわめて重要な地域といわれている(伊藤,1998)。極相群集は常緑広葉樹林であり､イタジイ(Castanopsis
sieboldii)を優占種とし、イスノキ(Distylium racemosum)、フカノキなどが樹冠を構成し､その下には密な中層木、下層木､地上植生の葉層をもっている。年平均気温22.4℃、年平均降雨量は2,036ｍｍである。次に､古見岳は標高470ｍで、西表島でもっとも高い。西表島は沖縄本島に次ぐ大きな面積をもち、島の約90％を亜熱帯の原生林で覆われ､1972年に国立公園に指定されている｡イリオモテヤマネコ(Felis
iriomotensis)をはじめとしてカンムリワシ(Spilornis chella perplexas)､セマルハコガメ(Cuora
flavomarginata evelynae)､キシノウエトカゲ(Eumeces kishinouyei)､仲間川や浦内川のマングローブ林、サキシマスオウノキ(Heritiera
littoralis)、ヤエヤマヤシ(Satakentia liukiuensis)など多くの天然記念物の指定を受けている。極相群集は常緑広葉樹林であり､イタジイやオキナワウラジロガシ、タブノキ(Perser
thunbergii)などを第一層とし、中層木、低層木、シダ類と重なり合った葉層のために､林内は暗い｡また、つる性植物のツルアダン(Freycinetia
formosana)が優勢で、いたるところの木の幹に絡み付いているのが見られる｡年平均気温は23.3℃､年平均降雨量は2,397ｍｍである。 Ⅹ．調査方法 沖縄島･西表島のそれぞれの原生的林分にプロットを設定して毎木調査を行った｡沖縄島は与那覇岳のイタジイ林､西表島は古見岳のオキナワウラジロガシ－イタジイ林を調査対象地とした｡プロットは､与那覇岳で66×４０ｍ設定し､内部は4×4ｍのグリッドセルに分割した。また、古見岳で50×70ｍ設定し､内部は１０×10ｍグリッドセルに分割した｡各調査地ともに、各グリッド単位で毎木調査を行った｡毎木調査は樹高２ｍ以上の生きている個体について標識し､樹種を記録してスチールメジャーによって胸高周囲長を、ノギスによって胸高直径(2方向)を計測した｡ ⅩⅠ． 解析方法 ＳＩＭＰＳＯＮの多様度指数 群集の種の豊富さ、多様性を求めるために、Ｓｉｍｏｓｏｎの多様度指数を用いた｡この式は2個体を取り出したときにそれが同種になる確率によって種類的な多様性を表現している｡Ｓｉｍｐｓｏｎの単純度指数：Ｄは以下の式となる｡
以上でｎｉ とＮ はｉ 番目の種の個体数と総個体数である。 なお、Ｓｉｍｐｓｏｎの単純度指数の逆数：1/Ｄは優占種の個体数の変化に鋭敏に反応するため､群集内の多様度の差を示すのに適しており､Ｓｉｍｐｓｏｎ多様度：1-Ｄは、サンプル数や種数に影響を受ないため、群集間の多様度の差を示すのに適している｡1/Ｄ，1-Ｄともに､値が大きければ大きいほど多様性は高くなるといえる｡ 重回帰分析 林分の垂直方向における多様度の推移が何の影響を受けているのかを知るために、重回帰分析を用いた｡重回帰分析は、一つの変数(独立変数)に対するほかの変数(従属変数)の関係を調べることを目的とし､3つ以上の変数間の関係を取り扱うときに用いる｡式は以下のように示される｡ Ｙ＝Ａ１X１＋Ａ２Ｘ２＋…＋ＡｐＸｐ＋Ａ０ 以上で従属変数：Ｙは結果を、独立変数：Ｘ１，Ｘ２，…，Ｘｐはそれに影響を与える原因を、回帰係数：Ａ１，Ａ２，…，Ａｐはその結果と原因の関係を示し、Ａ＞0で大きい値ほど、ＸはＹに正の影響を与え､Ａ＜0で小さい値ほど、ＸはＹに負の影響を与えているといえる｡本研究では､各調査地において従属変数に多様度の推移を､独立変数にＤＢＨ値､個体数､種数､林冠種の頻度等、それぞれの推移を当てはめ､回帰係数の値を求めた｡なお、回帰係数の値は有意水準Ｐ＜0.05の値を参考にした｡ コルモゴロフ－スミルノフ(Ｋｏｌｍｏｇｏｒｏｖ－Ｓｍｉｒｎｏｖ)の検定法 (Ｋ－Ｓテスト) 各調査地のイタジイの集団が異なるサイズ構造をもつかどうかを調べるために､コルモゴロフ－スミルノフ(Ｋｏｌｍｏｇｏｒｏｖ－Ｓｍｉｒｎｏｖ)の検定法を用いた｡この検定法は､二つの標本から相対累積頻度の差の絶対差d を求め､その中で最大絶対差Ｄ
が、Ｄ に対する棄却値Ｄα よりも大きくなれば､異なった分布を持つ標本であるといえる｡相対累積頻度の差の絶対差ｄ
および棄却値Ｄα は以下の式となる｡
α＝0.05に対しては、Ｋ.05＝1.35810 α＝0.01に対しては､Ｋ.01＝1.62762 以上でＦ は累積頻度分布の値を､ｎ は標本の大きさを示している｡ クラスター分析 植物個体群の様々な特徴から種がどのように分類されるか知るために、クラスター分析を用いた。クラスター分析は､異質な物が混じっている対象の中から､似たもの同士を集め､クラスターを作ることで､対象の分類を行う｡それらの結合距離が小さいほど、類似性が高いということが言える｡本研究では､頻度､萌芽率､最大直径､歪度､変動係数、相対現存量に基づいて各調査地の２０種を対象に分類を行った｡ 森下のＩδ指数 植物個体群の分布様式を求めるために､森下のＩδ指数を用いた｡Ｉδの値は個体群の密度との関係を表しており、以下の式から求められる｡
以上でｑ は区画数を､ｘi はi番目の区画内の種の個体数を、T は種の総個体数を示す。なお、Ｉδ の値が１よりも大きければ集中分布を、１よりも小さければ一様(ランダム)分布をする。 相関係数 植物個体群の分布が他の個体群の分布と関係しているのかどうかを調べるために、各種の分布の相関係数を求めた｡相関係数は､２変量xとyの関係を数量化したもので､相関係数：ｒは以下の式から求められる｡
以上でＸ，Ｙ はx，ｙ の平均を、Ｓｘ，Ｓｙ
はx，ｙ の標準偏差を示す｡相関係数ｒ はつねに －１≦ｒ≦１ の間の値をとり､ r≒－１ に対しては強い負の相関 r＜0 に対しては負の相関 r≒0 に対しては無相関 r＞0 に対しては正の相関 r ≒１ に対しては強い正の相関 を示している。本研究では､各調査地の２０種を対象とし、グリッドごとの分布相関を求めた｡有意水準はα＝0.05とし、スピアマンの順位相関検定より優位限界値： ｒＮ（α） よりも大きかった値を参考にしている｡ ⅩⅡ． 結果 （１）各調査地の植生 各調査地の個体数密度は、与那覇岳(沖縄島)9725/ha、古見岳(西表島)10830/ha、出現種は、与那覇岳265/ha、古見岳220/haであった。なお、各調査地に共通する種は４５種あった。また、本数割合は、与那覇岳では､イタジイ１０％、イスノキ、ボチョウジ､ヒメサザンカ(Camellia
lutchuensis)それぞれ6%、シシアクチ5%、オオシイバモチ､ホソバタブそれぞれ4%、その他63種が３～0.04%の範囲にあり､古見岳ではシシアクチ、ボチョウジそれぞれ10%、モクタチバナ(Ardisia
sieboldii)､イスノキ7%、オキナワウラジロガシ6%、シマミサオ､イタジイ5%、その他70種が3~0.03%の範囲にあった｡林冠を形成する種である与那覇岳のイタジイとイスノキ、古見岳のイタジイとオキナワウラジロガシは、それぞれ個体数および現存量において優占していた。最大幹直径は、与那覇岳ではイタジイの83.50ｃｍ、古見岳ではタブノキの77.48cmがあった。（表１－ａ，－b，表２） （２）種多様性 各調査地の多様性の高さをSＩＭＰＳＯＮの単純度指数の逆数（1/Ｄ）、SＩＭＰＳＯＮの多様度（1-Ｄ）を用いて求めると、与那覇岳(沖縄島)では1/D=26.5、1-D=0.96、古見岳(西表島)では1/D=22.3、1-D=0.96という値となった(図４)。 各調査地の全個体をDBH値１cmごとに分け、階層における多様度の推移を解析すると、与那覇岳ではＤＢＨ１０ｃｍ階で1/Ｄ＝40と高い多様度を示すが、他の階では変動は小さく、ＤＢＨ＞１４ｃｍでは徐々に低くなっていく傾向が見られた。一方、西表島では多様度の変動は小さく、ＤＢＨ＞２０ｃｍで徐々に低くなっていく傾向が見られた（図６）。多様度の推移に影響を与える要因を重回帰分析を用いて調べると、与那覇岳の多様度の推移はイタジイの個体数の割合に大きく影響を受けていることがわかった（Ｐ＜0.05）(表3)。西表島では優位な影響を与えている変数は見つからなかった。なお、各調査地の比較を行うと、ＤＢＨ＜15ｃｍまではほとんど等しく､ＤＢＨ＞１５ｃｍ（高木層）では古見岳の多様度が与那覇岳よりも高いということがわかった(図７)｡ （３）優占種のサイズ構造 各調査地の優占種について階層構造を見ると、各調査地ともに萌芽率の高いイタジイは二層化していた(図８-ａ-ｂ)。この二層化の境界は、与那覇岳(沖縄島)はＤＢＨ10ｃｍ前後、古見岳(西表島)はＤＢＨ１５ｃｍ前後というように古見岳が大きなサイズで二層化していた。同様に、上層の頭打ちになっているサイズも古見岳が大きかった。なお､KSテストより各調査地のイタジイのサイズ構造には有意な差があることがわかっている（最大絶対差Ｄ＝0.18は、Ｄの棄却値Ｄα＝0.16（α＝0.01）よりも大きい）。また、与那覇岳のイスノキはDBH35cm前後で頭打ちになっており、古見岳のオキナワウラジロガシは指数関数的にDBHが大きくなっていた(図８-ｃ，-ｄ)。 （４）ヒストグラム 個体数の多い種から各調査地に共通する種をあげ、ヒストグラムを用いてサイズ構造を解析すると、ヒストグラムの型は二つのタイプに分かれた（図９-ａ，-ｂ）。一つは、L字型分布を示す種；ホソバタブ、オオシイバモチ、イスノキ、エゴノキ(Styrax japonica)、イタジイ(与那覇岳)、オキナワウラジロガシ(古見岳)、次に、パルス型分布を示す種；フカノキ､モクタチバナ､コバンモチ(Elaeocarpus
japonicus)､イタジイ(古見岳)に分かれた。なお、これらのサイズ構造の型はイタジイ､イスノキ以外は与那覇岳と古見岳で違いがあまり見られなかった。 （５）クラスター解析 各種の頻度や萌芽率､DBH値の最大値､歪度､変動係数、現存量などを変数とし、クラスター解析を行ったところ､古見岳では低木のタイプがはっきりとクラス分けされてあらわれたのに対し､与那覇岳では低木のクラスと中高木のクラスとの境界がはっきりとしていなかった｡また、各調査地ともに高木の種ははっきりとクラス分けされていた(図１０-ａ，-ｂ)。 （６）Iδ指数 水平分布について各種の分散の具合をＩδ指数を用いて解析を行い、コドラートの大きさを４×４～１６×１６(m)へと変えたときのＩδ値の推移を比較した(図１１-ａ，-ｂ，-ｃ)｡ただし､古見岳(西表島)では、コドラートの大きさが１０×１０(m)よりも小さくできなかったため、与那覇岳(沖縄島)の値を用いている｡これより、種は大きく二つのタイプに分けることができた｡一つは､コドラートが小さいとき(4×4)Ｉδ値が高くなる種で､集中分布をする種；エゴノキ､ホソバタブ､タブノキ、ヤマヒハツ(Antidesma
japonicum)､コバンモチ､ハクサンボク(Viburnum japonicum)､イタジイ(図１１-ａ)、もう一つは、コドラートの大きさが変わってもＩδ値はあまり変わらない種で、ランダム分布をする種；ボチョウジ､イスノキ、オオシイバモチ、シマミサオノキ、ヒメサザンカ(Camellia
lutchuensis)､ムッチャガラ(Ilex maximowicziana)､ヤンバルマユミ(Euonymus
tashiroi) (図11-ｃ)であった｡また、その中間の種も見られた(図11-ｂ)｡ （７）種間の分布相関 各調査地の種間の分布相関をスピアマンの順位相関係数より分析すると、同じ階層で正の相関が少ないという傾向が見られた(表4-ａ，-ｂ)｡また､同じ階層で正の相関が見られた場合､与那覇岳のシマミサオノキとヤマヒハツ､古見岳のオオシイバモチとエゴノキ､シシアクチとヤマヒハツなどは、水平方向の分布様式（集中分布､ランダム分布）を違えている傾向が見られた。 | ||||||||||||||||||||||
