全 文 ：第 26卷第 8期
2006年 8月
生 态 学 报
ACTA ECOLOGICA SINICA
Vo1．26，No．8
Aug．，2006
金沙江干热河谷山地植被恢复区土壤
种子库和地上植被研究
罗 辉 ，王克勤
(西南林学院环境科学与工程系，昆明 650224)
摘要：土壤种子库在植物种群动态中起着重要作用。土壤种子库可缓解种群的灭绝过程，保存群落中植物种的表现特征，是植
被天然更新的物质基础。通过对金沙江干热河谷山地植被恢复区(包括水平阶、自然坡面、沟底)和未恢复区(包括放牧地)的土
壤种子库和地 上植被的组成、大小及多样性进行 比较研究表 明，植被恢复 区土壤种子库和地上植被 的密度 、丰富度 、多样性及 均
匀度均大于未恢复区。恢复区地上生物量要远大于未恢复区。水平阶和各类型间的土壤种子库密度与地上植被密度差异显
著。土壤种子库 中草本植物 占很大 比例 。孔颖草和扭 黄茅 是土壤种 子库和地上植被的两大优势种 ，两者 的个 体数量 、重要值及
生物量最大。土壤种子库和地上植被有较高的相似性 ，且随着恢复程度的加深 ，相似性有增高的趋势；土壤种子库密度和地上
植 被密度之 间关 系可 以用二次 和三 次曲线拟 合。
关键词 ：土壤种子库 ；tg上植被 ；物种多样性 ；干热河谷
文章 编号 ：1000—0933(2006)08—2432—11 中图分类号：Q948 文献标识码 ：A
Soil seed bank and aboveground vegetation in Jinshajing Hot—Dry River Valley
Hillslope vegetation restoration site
LUO Hui，WANG Ke—Qin (D 九 眦 E加 r0n 0z sc 删 口 Engineering，Southwest For~try c0z ，Kunming 650224，c^ )．A嘲
Ecologica Sinica，20O6，26(s)：2432—2442．
Abstract：Soil seed bank plays an important role in the compo sition of different plant communities and especialy in their
conservation．Although Soil seed bank，aboveground vegetation and their relationship have been the subject of much recent
atention，litle is known about the size and species composition of soil seed bank and aboveground vegetation in semi—arid hillslope
grasslands and understanding of how these compo nents interact to determine the impo rtance of seed banks to regeneration is
limited．We assessed the size and species compo sition of a soil seed bank and aboveground vegetation in an experiment with 36
vegetation quadrats and 108 soil samples in terrace，slope，guly and grazing land that represent a range of habitats within a
hilslope grassland in Jinshajing hot—dry river valley of Yunnan．Terrace，slope and guly represent restored site and grazing land
typifes unrestored site．We identifed 21 taxa in the seed bank with a median of 7 species／me and a median density of 5498 seeds／
m
2
， while in aboveground vegetation． 1 9 species were observed with a median of 6 species／m2 and a median density of 1088
plants／me
． Both seed bank density and aboveground vegetation density among grazing land，guly，slope and terrace difered
signifcantly．There was an absolutely high proportion of herbaceous species in the seed bank and aboveground vegetation．
Gramineae predominated over both seed bank and vegetation．The most frequent seeds and plants were Bothriochloa peausa(L．)
基金项目：国家 自然科学基金资助项 目 (30170779)；云南省自然科学基金重点资助项 目(2001D0008Z)
收稿 日期：2005—05—29；修订日期：2006—03—15
作者简介：罗辉(1979～)，男，硕士，主要从事恢复生态学研究．E—mail：roadways@126．com
*通讯作者 Coresponding au~or．E-mail：kqwang@public．km．yn．cn；kqwang@swfc．edu．cn
Foundation item：The project was supported by The National Natural Science Foundation of China(No．30170779)；The Natural Science Foundation of Yunnan
Province，China(No．2001D0008Z)
Received date：2005—05—29：Accepted date ：2006—03—15
Biography：LUO Hui，Master．mainly engaged in restoring ecology．E—mail：roadways@ 126．COB
维普资讯 http://www.cqvip.com
8期 罗辉 等 ：金沙江干热河谷山地植被恢复区土壤种子库 和地上植被研究 2433
A．Camus and Heterepogon contortus(L．)Beauv that had the highest individual number，importance value and biomass．In the
seed bank，the seeds of Bothriochloapertusa(L．)A．Camus and Hetempogon contortus(L．)Beauv accounted for 50．68％ and
f0r 33．10％ of the total seeds respectively．In aboveground vegetation，the individual number of Bothriochloa penusa(L．)A．
Camus accounted for 55．66％ of the total and Hetempogon contortus(L．)Beauv accounted for 29．86％ ．The biomass of
Bothriochloapenusa(L．)A．Camus and Heterepogon contortus(L．)Beauv．accounted for more than 70％ of total，reaching
206．71 m and 147．76 g／m2 respectively
． Bothriochloapertusa(L．)A．Camus and Heterepogon contortus(L．)Beauv had the
highest importance value of 193．01 and 159．99 respectively．
Density，biomass，species richness，spe cies diversity and evenness were the highest in terace while those in grazing lan d
were the lowest．Similarities between the seed ban k and aboveground vegetation were moderately high and not very different among
slope ，guly and terace，except for grazing land，tending to increase when restorative stage progressed．Th is result contrasts with
some other studies where the seed bank contributes very litle to the seedling flora and vegetative growth clearly overwhelms sexual
reproduction．Th e hypothesis about signifcant functional correlation be tween soil seed bank density and abovegreund vegetation
density is conformed．Correlation between soil sed bank density and aboveground vegetation density can be described as quadratic
and cubic curves．Th e strong similarity be tween vegetation and the seed ba nk is atributed to the great proportion of the species
Bothriochloa pertusa(L．)A．Camus and Heterepogon cortortus(L．)Beauv．that are seed—prefusive and whose seeds have a
signifcant viability in the ground．Th e high density，biomass，species richness，spe cies diversity and evenness of the reclaimed
site is related to the sufficiency of heat and water supplies for spe cies establishment and growing in the site，which partly reflects
our efective eforts on the hilslope grassland restoration． We believe that our vegetation restoration eforts have altered the
micrehabitat conditions of the site and provided a favorable habitat for species to establish and grow．
Key words：soil seed ba nk；aboveground vegetation；species diversity；hot—dry river valey
1 Introduction
Soil seed bank refers to the ungerminated but viable seeds that lie in the soil ．Many plant species have the capacity
to produce seeds that remain dormant in the soil for several years to several decades．In most habitats，ranging from the
arctic to the tropics，deserts to wetlands，natural to cultivated lands investigators have found evidence for seed banks．Seed
banks are thought to be ecologically and evolutionarily important in the dynamics of plant populations．Seed banks can
function as reservoirs of genes and／or gene complexes．This means that seed persistence，the carry—over of viable seeds in
the soil for multiple years，can buffer the effect of local extinction of genotypes in the non—dorm ant portion of the
population，and act to maintain genetic variation，during periods when seedling do not survive to become reproductive
adults．Seed banks can also buffer a population from extinction and preserve the representation of a plant species within a
communitv|2
．
Seed banks and their relationships to vegetation have been the subject of much recent attention．The understanding of
the potential of a seed bank to alter vegetation composition，its potential for restoring richness in species and maintaining
floristic diversity are some of the reasons that have motivated researchers to compare the composition of the aboveground
vegetation with seed reserves hidden in the soil．The seed bank is a major functional compartment of a plant community，in
that its role as a storage compartment allows population maintenance according to changes in reproduction perform ances
either between years or between sites ．Thompson & Grime defined four types of seed banks among the most common
species in temperate regions，characterized by singularities in the persistence of their seeds in soil ．These types range
from transient seed banks constituted by seeds that germinate in greater numbers immediately after dispersal，to persistent
seed banks with seeds that remain dormant in the soil over a longer period(more than 1 year)until environmental and／or
temporal conditions are favorable．Seed banks of trees，shrubs and shade—tolerant herbaceous species in temperate
woodlands are generally transient．while shade—intolerant species are the primary constituents of persistent seed banks ．
维普资讯 http://www.cqvip.com
2434 生 态 学 报 26卷
However，most past studies of seed banks and their relationships to aboveground vegetation have been concentrated on
arable lands， alluvial wetlan ds， and sucessional forest where seed bank composition reflects historical lan d use
． While
fewer investigation of seed ban ks in semi—arid waste hillslope grassland have been made，it is clear that seed banks are
important in waste grassland recovery after disturbance．If we aim to regenerate the local vegetation with a certain
modifcation，the ecological function of seed banks in semi—arid hillslope grassland must be fully understood
．
In this study，we characterized the soil seed bank and aboveground vegetation in a 10．hectare tract of previously
abandoned hilslope grassland that is without precise boundaries but surrounded by agricultural land in Jinshajing hot．dry
river valley．We sampled many diverse habitats within the tract of the grassland
． With this approach we were able to
explore relationships among the existing vegetation and seed bank composition，within a single，continuous，regenerating，
hillslope grassland ecosystem．Our general objectives were：(i)to determine the size，composition，species richness and
species diversity of the seed bank and aboveground vegetation in a broad range of habitats within this large fragment of
grassland， (ii)to describe the relationships between the seed bank and existing vegetation within this grassland，and(iii)
to exami ne whether and to what extent the community structure an d productivity had been optimized by our facilitating
restoration efforts．
2 M aterials and M ethotis
2．1 Study site
The present study was affiliated with the project of Mechanics and Application of Micro．Catchment Water Harvesting
Agriculture and Forestry in Jinshajing River。which was initiated in 2002．It was carried out in the hilly land of Laocheng
Village，Yuanmou County(25。43 52”North，101。51 03”East；altitude：1100— 1200 m)in the north．central Dart 0f
Yunnan Province(1ocated≈ 190 km northwest of Kunming City)。which is a typical hot—dry fiver valley area in southwest
China．
This region is altematively affected by tropical monsoon and foehn from the Indian Ocean．Under the frequent
domination of foehn，the region is characterized by the aridity and hotness of weather．Moreover Foehn winds cause a hot
dry season of seven months，and scorching dry spells afflict the wet seasons．Ecologically the areas are simlar to tropical
savannas and called semi‘natural savanna or secondary savanna ．Th e mean annual temperature is 21．7 oC ranging from
14．9 oC in December to 27．1 oC in May，and the mean annual precipitation totals 629 mm (most ofwhich is rainfal1)．Th e
mean annual evaporation，which is nearly 6 times the precipitation，averages 3729 mm．Rainfall is not distributed evenly
throughout the year，and most rain falls during June to October．Th e 7-month—long dry season，from Novemb er to May，
averages less than 100 mm or 14％ of the mean annual rainfall(Fig．1)．
The dry climate makes successful tree planting a challenging operation．It also makes natural regeneration extremely
dificult after forests are harvested or cleared．The survival rate of trees planted using planting technologies that are
effective in other areas was very low in this area during the period 1952— 1988．Only 6％ of Yuanmou’s land area was
covered with forests in 1993．Th e majority of its designated forestland is covered with por quality bushes or grass and
sometimes there is just bare land．In the short wet season，severe soil erosion and serious land degradation resulting from
excessive exploitation and unsustainable practices haunts the area．To address this problem，restoration of cover by
planting trees，shrubs and grasses is necessary ． Although it is very challengeable for f0resters／ecolo sts to improve the
deteriorating ecological environment and achieve economically and ecologically sustainable development of this ailing
region，many researches have been successfuly undertaken and the vegetation in certain area has been partly restored．
2．2 Sampling design
Th e experiments were conducted in four contrasted vegetation types．Th e first is a frequently mowed and grazed
agricultural field(called grazing land)．Th is field remaining as a control tool epitomizing unrestored plot is buting against
维普资讯 http://www.cqvip.com
8期 罗辉 等：金沙江干热河谷山地植被恢复区土壤种子库和地上植被研究 2435
暑
暑
一
宝
highly restored plots．Except for the controlled field，the Month
。thers are all within the expe mental area·Three 40m l。ng，
Fig．1 c1i眦te dia 瑚 aI1d m跚 m。nthly evap0mtion(t叩 。f 6gure)f0r
nonparallel transects with diferent bearings were selected
study sit。
within each vegetation type． Each transect comprised 3
regnlarly spaced quadrats(1 m×l m，1 0m apart)and epitomized the vegetation around⋯．Altogether，1 2 transects and
36quadrats were set up．
2．3 Seed bank sampling
The seed bank was sampled in January 2004。At that time，seeds released the previous summer had a natural cold
stratifcation，seeds from the 2004 season had not been released，and there had not been an opportunity for the seeds to
germinate that spring．Th ere is some disagreement in the literature on the best time to sample seed banks；we recognize
that we are sampling both persistent and transient seed banks ．We used the method developed by Lavorel et a1．
consisting of a randomly sampling within each sampling quadrat(three replicates per quadrate，each replicate 30 cm
apart) ．Altogether 108 soil samples were colected(9 per transect；27 per vegetation type)．Soil samples were taken in
form of a square (10 cm×10 cm)along a central transect running the length of each quadrat devoted to seed bank
sampling．The majority of viable seeds is normally concentrated in the very first centimeters of the ground(i．e．the litter—
fermentation．humus layers)，with fewer seeds found out to 10 cm ．The depth of sampling was 5 cm，to yield a total
volume of 0．5 L and a total soil surface area of 0．03 m2 per quadrat．Th e soil samples were then sorted to elim nate the
plant fragments and stones and kept in ventilating bags．
2．4 Seedling emergence technique
Th e density and composition of the seed ban k were determined by observing seedling emergence． We chose the
“emergence method” for analysis of the seed bank，because our primary goal was to determ ine the abundance and
distribution of viable seeds that could germinate under field conditions． Gross showed that elutriation， by including
nonviable seeds．gave higher estimates of seed density than emergence ．Poiani and Johnson found that the“emergence
method”gave an accurate assessment of the number of species and the relative abundance of seeds present compared with
the actual identification of seeds ． Although it is generally accepted that the “emergence method’ gives biased
assessments of the seed bank because greenhouse conditions ale never exactly the same as field conditions ‘ ，it is the
most appropriate method for measuring the seed bank composition and exploring relationships between the seed bank and
aboveground vegetation．
The seedling emergence study began in spring(March)．The soil samples were placed in a wal31 greenhouse where
temperatures are similar to those outdoors in the Yunmou area and they were kept moist．Each sample was divided in half
and spread out in two germination flowerpots(9130cm×90 cm)to the depth of 2 cm over 4 cm of sand(Th e sand was
previously sterilized and put into the pots to let exogenous seeds in it fully germinate)．Each pot from each plot was
维普资讯 http://www.cqvip.com
2436 生 态 学 报 26卷
assigned a random position in one of two replicates．Al pots were watered as often as needed to keep the soil moist．We
fertilized the pots with standard plant food every 2 weeks to encourage faster growth and avoid seedling death[13]
． Once the
first seedlings appeared observations and identifcations were facilitated using seedling floras．Emerging seedlings were
identifed，recorded and removed weekly or transplanted to grow to maturity for later identifcation．In order to favor the
．
maximum of germination，seedlings were pulled out after identifcation to maintain a low density in the germination pots
and to allow the germination of other seeds．At the end of the first two months of the experiments，the soil samples were
carefully turned over in order to facilitate the emergence of new seedlings ．After six months， the sampling was
terminated as no more emergence occurred for several consecutive weeks．To quantify germination of ambient seeds blown
into the experimental pots inside the greenhouse，several control pots containing a 2cm layer of sterile weed-free poting soil
over 4 cm of sand were placed randomly in each replicate block．We found no signifcant diferences in the numb er or
density of species between replicate blocks，so the results were comb ined for all analyses．
2．5 Vegetation characterization
To record every species of vascular and herbaceous plant present within the whole study plots，we ma de surveys in
mid-autumns of 2003 when the development of plants was optima1 ．Floristic information including taxa，percent coverage，
individual numb er an d biomass was obtained by surveys carried out in each quadrat．We used the nomenclature of Flora
Sinicae(Delectis Flora Sinicae Agendae，1959— 1999)．After visually estima ting taxa，percent coverage and counting the
individuals，we cut down all the individuals wi thin the qudrats for further determining the biomass of each tax on．Finally
plant samples of each taxon within each quadrat were dried completely in an oven and then weighed．
2．6 Data analyses
We analyzed two diferent aspects of seed bank：composition and properties．Composition is simply the numb er of
seedlings of each kind at each quadrate．To characterize the seed bank of each vegetation type as a whole，we calculated
the mean species richness(S)，mean seed density(seeds／m soil surface)，the proportion of taxa，and the mean species
diversity from all the quadrats of the same vegetation type．W e used the Shannon-Wiener index，the Simpson index and
Hurlbert’s probability of intraspecific encounter to indicate species diversity，and the Pielou index to indicate species
evenness
[ 一
． We also calculated these same characteristic properties for aboveground vegetation of each type after
determining relative abundance(RA)，relative coverage(RC)，relative frequency(RF)and importance value(IV)of
each species．
Simpson diversity index： D=1一∑P ；
Shann。n-Wiener diversity index： H =一∑Pi ln P；
Probability of intraspecific encounter：
Pielou evenness index：
PIE=∑[(N ／N)(N—N；)／(N一1)]；
．，= D／(1—1／．s)．
where S = the species richness， N = the individual numb er of the ith species， N = the individual numb er of all
species，and P = the proportion of the ith species．
Soil seed bank density，aboveground vegetation density and soil property among different types of vegetation were both
compared by means of a one-way analysis of variance(ANOVA)followed by Tukey’s multiple comparison test at P =
0．05．All analyses of variance were conducted with MINITAB Release 14．13．(Minitab Inc．，1972— 2004)．Before
performing ANOVA all data was calculated as a mean±stan dard deviation an d rounded to the nearest whole numb er．
The similarity between species composition in the seed bank and in the aboveground vegetation was assessed using
Jaccard’s similarity coeficient based on species presence and absence ．Th is coefficient was calculated between all types
of vegetation and seed bank．
维普资讯 http://www.cqvip.com
8期 罗辉 等 ：金沙 江干热河谷 山地植 被恢 复区土壤种子库和地上植被研究 2437
C．= a／(a+b+c)
where a = the number of species common in sample A and sample B，b = the numb er of species present in A but
absent in B，c = the numb er of species present in B，absent in A．The result，as measured by the Jaccard’s
coefficient，ranged between 0，for no species shared in common，and 1，for complete concurence．
We assessed the relationships between the density of the seed bank and the aboveground vegetation，which is possible
by using a regression analysis．Th e data of density was logarithmically transformed before the statistical analysis to meet the
assumption of norm ality and homogeneous variances and reduce positive skew[20
．
3 ResulIs
3．1 Overall richness and density of the soil seed bank and aboveground vegetation
A total of 1980 seedlings(265 in grazing land，392 in guly，447 in slope，876 in terrace)and 21 species(16
species in grazing land，1 3 species in guly，15 species in slope；17 species in terace)were observed during the trial
(Table 1)．The total seed numb er varied significantly among the samples．The density was signifcantly higher(P<
0．001)in terrace(9738 seeds／m )than in grazing lard(2941 seeds／m )，gully(4369 seeds／m )and slope(4942seeds／
m )．Th e numb er of seed bank taxa per quadrat ranged from 4— 15 with a median of 7 species，while seed density in
individual quadrats ranged from 2464 to 1 1605 seeds／m with a median of 5498 seeds／m overal1．A total of 19 species was
identifed in the aboveground vegetation(8 species in grazing land，14 species in guly，16 species in slope；18 species in
terace)ranging from 3 to 1 1 per quadrat with a median of 6，while vegetation density in individual quadrats ranged from
606 plants／m (in grazing land)to 1904 plants／m (in terace)with a median of 1088 plants／m (Table 2)．
3．2 Composition of the soil seed bank and aboveground vegetation
Th e soil seed bank was dominated by only a few species：91．56％ of the seedlings came from 3 shade—intolerant
graminoid species occuring in more than 95％ of the quadrats：Bothriochloa pertusa (L．)A．Camus(50．68％)，
Heteropogon conto~us(L．)Beauv．(33．10％)and Eulaliopsis binata (Retz．)C．E．Hubb．(7．78％)．Seeds of
Bothriochloa pertusa(L．)A．Camus were the most numerous overal(Table 1)．4 species(Acacia confusa Merr．，
Eupatorium adenophorum L．，Datura stramonium L．，Eucalyptus camaldulensis Dehn—hardt．)were only sampled in
grazing land and 2 species(Atylosia mollis(Willd．)Benth．，Phyllanthus urinaria L．)were observed in gully，slope and
terace except in grazing land．Most species we found in the aboveground vegetation had almost emerged．Notably absent
from the seed bank were shade．tolerant herbs and some semi—shrubs(Taraxacum mongolicum Hand．一Mazz．，Stellera
chamaejaslTte L．，Desmodium podocarpum ssp．oxyphylum DC．)，which are found in a moderately high numb er in the
aboveground vegetation in the hot—dry valley．There was an high proportion(98．53％)of herbaceous species in the seed
bank and with the exception of Albizzia molis(Wal1．)Boiv．，we found relatively few tree seeds．
Like the soil seed bank the aboveground vegetation was overrepresented (95．92％)by the 3 graminoid species：
Bothriochloa pertusa(L．)A．Camus(55．66％)，Heteropogon contortus (L．)Beauv．(29．86％)and Eulaliopsis binata
(Retz．)C．E．Hubb．(10．40％)．Besides these 3 dominant species，Cynodon dactylon(L．)Pers．，Atylosia mollis
(Willd．)Benth．，Albizzia mollis(Wal1．)Boiv．and Dodonaea viscosa(L．)Jacq．occured in all types of vegetation．
0f 21 taxa identified in the seed bank，5(Acacia confusa Mer．，Eupatorium adenophorum L．，Datura stramonium L．，
Eucalyptus camaldulensis Dehn—hardt．，Oxalis corniculata L．)were not found in the vegetation samples of any of the 36
quadrats(Table 2)．These 5 alien species are likely to have been blown or carried in from the adjacent modified landscape
or from locally disturbed intersection between the abandoned field and the agricultural land．As mentioned above，except
for Oxalis corniculata L．the other 4 were only observed in grazing land that was most likely exposed to unexpectab le
human and animal intrusions．There was a little higher proportion(99．05％ )of herbaceous species in the aboveground
vegetation than in the seed bank，which can be attributed to some herbaceous species found in the vegetation that were not
维普资讯 http://www.cqvip.com
2438 生 态 学 报 26卷
Note：Values are mean
P >0．05：A：Annual herb B
4-standard deviation；Values with diferent leters mean signifcant diference(P<0．05 )：Lf=Life form。NS=Not signifcant at
Biennial herb P：Perennial herb W ：Wody species(shrub or tree)
Table 2 Characteristics of aboveground vegetation(Mean±SD。plants／m2)
维普资讯 http://www.cqvip.com
8期 罗辉 等：金沙江干热河谷山地植被恢复区土壤种子库和地上植被研究 2439
3．3 Properties of the soil seed bank and aboveground vegetation
Both the soil seed bank and aboveground vegetation included mostly herbaceous taxa，and Gramineae followed by
Leguminosae and Compositae predominated over both seed bank and vegetation．The analysis of variance showed that the
differences of density among grazing land，gully，slope，and terrace were significant in the seed bank (Table 1)and
aboveground vegetation(Table 2)．Density was much higher in terace than in grazing land in both seed bank samples and
in the vegetation quadrats，and Tukey’s multiple comparison test showed that the diference of overall density of soil seed
bank and aboveground vegetation between terrace and other types was significant．
For the aboveground vegetation as a whole，Bothriochloa pertusa(L．)A．Camus had the highest importance value of
193．01．Heteropogon contortus (L．)Beauv．was less copious than Bothriochloa pertusa (L．)A．Camus with the
importance value of 1 59．99．Stellera chamaejasme L．，which sporadically occurred in sunny teraces or slope，was at the
botom of the list with the lowest importance value of 2．86．Desmodium podocarpum ssp． oxyphyUum DC．， Eclipta
prostrata L．，Solanum xanthocarpum Schrad．et Wend1．，Xanthium sibiricum Patrin．，Taraxacum mongolicum Hand．一
Mazz．and Sida szechuensis Matsuda were less important with values under 20(Table 3)．
Th ere were significant diferences of biomass among diferent vegetation types．Th e biomass of the plants growing in
the terace was the highest，which reflects the remarkable
productivity in highly restored sites．Th is mi ght be related
to the favorable heat and water conditions within the terace
where our assistance to facilitate the plants’growth was
applied． Th e biomass in grazing land was the lowest
because of the iregu larly heavy grazing and mowi ng．Th e
biomass in slope and gu lly was intermediate for the
insuficiency of heat and water． The biomass of different
plants varied greatly even in the same habitat．W ith more
than 70％ of the total biomass Bothriochloa pertusa (L．)
A．Camus and Heteropogon contortus (L．) Beauv．
domi nated the whole vegetation， while other species
contributed relatively litle biomass to vegetation． The
biomass of these two species reached 206．71 g／m and 147．
76 g／m respectively(Table 4)．
Species richness in terrace was the greatest both in the
soil seed ban k and in aboveground vegetation． Richness
also tended to be greater in aboveground vegetation than in
the soil seed bank， although not signifcantly．Similarly，
higher richness in plant in the seed bank was linked to a
higher one in vegetation．Terace followed by guly had the
highest diversity 0 H & PlE both in the seed bank and in
vegetation， while grazing field had the lowest． In seed
bank，evenness or relative distribution of individuals among
species was greater in terace and gu lly． However， in
aboveground vegetation the individuals in gully and grazing
land had a higher evenness(Table 5)．
Table 3 Im portance value of alI species in aboveground vegetation
Note： RA = Relative abundance， RC = Relative coverage，RF
Relative frequency，IV = Importance value
Table 4 Biomass in aboveground vegetation(g，m )
维普资讯 http://www.cqvip.com
2440 生 态 学 报 26卷
3．4 Similarity and relationship between the soil seed bank and aboveground vegetation
Sim larity between soil seed bank and aboveground vegetation at individual quadrats was high：values of Jaccard’s
coefficient of community ranged from 0．412 to 0．842(with a mean of 0．719)and tended to increase when the restorative
stage progressed(Table 6：sv)．Fluctuating similarities among different types were observed both in vegetation quadrats
and in seed bank samples：Simi larity among different types of aboveground vegetation varied between 0．412 and 0．790
(Table 6：v)and similarity among diferent type of seed bank ranged from 0．526 to 0．790(Table 6：s)；The degree of
sim larity in seed bank was moderately higher than that in vegetation quadrats，which reflects less fluctuation in seed bank
in successional process．In genera1 ．the similarities were ab out 0．649 among diferent types of vegetation in seed bank an d
ab oveground vegetation，revealing a relatively strong simi larity in most of the cases．
Regression an alysis showed there was moderately strong positive correlation between soil seed bank density and
aboveground vegetation density．Quadratic and cubic curves are suitable for describing the relationship between soil seed
bank and aboveground vegetation in density(Fig．2)．Our result agreed with that one of the research conducted by O’
Connor& Picket，Zhao et a1．and Zhao et a1．during which significant corelation was found，but difered from the
Th ompson & Grime study during which they f0und weak corelation between soil seed bank and aboveground
vegetation【 ’ 一
．
4 DiSCll~sion
4．1 Seed bank and aboveground vegetation similarity
Sim larity between soil seed bank and aboveground ve—
getation has long been a contention for many researchers．
Most studies of grasslands predominated by perennial grass—
es have found few similarities between the seed bank and
the vegetation．Th ese discrepancies have been explained by
the minor contribution of the dominant perennial meadow
species to the formation of seed banks．These species gen—
erally have a low seed production because they alternate
sexual reproduction with vegetative form s and their seeds
have a short—term persistence in the soil ． Moreover，
where important seed banks can be found in such perennial
grasslands，the soil seed banks often contain large numbers
of seeds of annual ruderal species(species‘R’sensu)that
reduce the sim larity between the vegetation and the seed
bankl 2 4_
．
In recent years， exceptions to the weak simi larities
between seed bank and vegetation have been found，wi th
c】ose sim 】arities as in freshwater tidal marshes．in annual
Table 6 Jaccard’s similarity coeficient among diferent types of soil
seed bank and aboveground vegetation
v)Similarity among diferent types of vegetation；s)Similarity among
diferent types of seed bank；sv)Similarity between seed bank and abeveground
vegetation of the same type
◆ Obse~ed ⋯ Quadratic 一 ·Cubic
Y=一2 166x +23 575x 一84 606x+ 103 06
R2=O 8602 P
3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4
log seed bank density(viable seeds／m )
Fig．2 Hypothetical relation between soil seed bank density and aboveground
vegetation density
3 2 ● O 9 8
3 3 3 3 2 2
E／slu旦d1三 c 口c0【】墨 眦 > 0一
维普资讯 http://www.cqvip.com
8期 罗辉 等：金沙江干热河谷山地植被恢复区土壤种子库和地上植被研究 244l
Mediterranean pastures or in a desert short grass community in New Mexico[ ～
． The common denominator of these
disturbed areas is the predominance of annual species that excessively produce seeds during growing season ．Even in
tropical forest domi nated by wooden species．strong simi larities are found ⋯ ．Although the seed bank has been studied in
many types of habitats，its functioning remains rather poorly known ．
Unlike most studies of temperate perennial grasslands，we do find a correspondence between the species composition
of the seed bank and the aboveground vegetation in this subtropic hilslope grassland．Th is accordance is much higher in
restored sites and tends to increase when the restorative stage progresses．Th e strong similarity between vegetation and seed
bank is attributed to the great proportion of perennial species(mainly Bothriochloa pertusa (L．)A．Camus and
Heteropogon contortus(L．)Beauv．)that are present in the vegetation and whose seeds have a signifcant viability in the
ground in relation to the strategies of opportunistic species．Ahhough these species are perennial yet they almost promote
sexual reproduction compared to the vegetative way and have a considerably high seed production as well as vegetative
growth，as could greatly contribute to increasing this simi larity．At the same time we also find factors that tend to cause the
dissimilarity between vegetation and seed ban k：factors particularly contributable to the dissimilarity between vegetation and
seed bank could be caused by species，which are only present in the vegetation but absent in the seed bank(described as
the disporum type，i．e．，species that show no evidence of forming a seed bank，sensu Thompson) ．In addition，some
wooden species identifed as having a persistent seed bank do not always emerge in the seed bank．The results show these
species were identifed only once or twice in the seed bank．It is thus inherently dificult to detect these rare species in the
seed bank．Although these factors incline to decrease the simi larity between soil seed bank and aboveground vegetation to
a certain extent，the predominance of seed—profusive perennial largely ofsets such inclination．
4．2 Comparison between restored and unrestored sites in seed bank and ab oveground vegetation
The structure and composition of the seed bank and aboveground vegetation in grazing land(unrestored plots)is
oversimplistic with low density，biomass，species richness，species diversity and evenness．This is inseparably related to
the hot and dry climate providing harsh growing conditions for plants and critical constraints(such as lack of water supply)
to vegetation establishment and can be partly attributed to the relative lack of abundance in seed bank．Lack of seed bank
in these unrestored plots may be a result of the combined efect of erosion，foraging animals and wind．However，compared
to grazing land，the reclaimed site(including slope，gully and terrace)represents relatively high density，biomass，
species richness，species diversity and evenness in the seed bank and aboveground vegetation．Th is is inextricably related
to the suficiency of heat and water supplies for species establishment and growing in the restored site，which partly reflects
our efective efforts on the hillslope grassland restoration．
We believe that our micro—catchment’s water harvesting system initiated in 1 998 has provided a favorable habitat for
species to establish and grow by reducing erosion，providing suficient water supply and improving edaphic conditions．
Moreover it alters microhabitat conditions of the site and litter accumulation promotes germination and seedling survival that
leads to the establishment of alien species．Evidence for this is the number and diversity of grass and shrubs species that
have volunteered onto the reclaimed site since the original foundation of the system．
5 Acknowledgment
We thank Q．B．Chen for providing generous advice and support in the design of the seed—bank sampling protocol，
G．S．Fan and J．M．Zeng provided critical support identifying seedlings，and the research could not have been
completed without them．We thank field assistant H．F．Deng for tireless support collecting and processing samples，H．
Li for careful assistance watering and fertilizing the seedlings，and J．F．Mo for facilitating the soil sample analyses．
References：
[1] Alaby M．Oxford Dictionary of Ecology(2nd edition)．Oxford：Oxford University Press，1998
维普资讯 http://www.cqvip.com
2442 生 态 学 报 26卷
[2] Kalisz S．Experimental determination of seed bank age structure in the winter annual Colinsia ．Ecology，1991，72：575—585．
[3] Touzard B，Amiaud B，Langlois E， et 02．The relationship between soil seed bank，aboveground vegetation and disturbances in an eutrophic aluvial
wetlan d of W estern France．F1ora，2002，197：175—185．
[4]
[5]
[6]
[7]
8
9
lO
ll
[12]
l3
14
l5
l6
l7
l8
l9
20
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[3O]
Th ompsen K，Grime J P．Seasonal variation in the seed bank of herbaceous species in ten contrasting habitats．Journal of Ecology ，1979，67：893～921．
Th ompson K．1The functional ecology of seed ban ks． In：Fenner M ed ． 1The ecology of regeneration in plan t communities．London： CAB Internationa1．
1992．23l一258．
Ou x K，Jin Z Z．A research on the vegetation types of the dry-hot fiver valey in Yuanmou．Acta Botanica Yunnanica，1987，87(3)：271～288．
Shu W S，Lan C Y，Wong M H，et 02．Early natural vegetation and soil seed banks in an abandoned quary．Acta Ecologlca Sinica，2003，23(7)：1305
～ l3l2．
Leckie S，VeUend M，Bell G，et 02．1The seed bank in an old-growth，temperate deciduous forest．Canadian Journal of Botany，2000，78：181～192．
Lavorel S，Debussche M，Lebreton J D，et 02．Seasonal patterns in the seed bank of Mediterranean oldtlelds．Oikos，1993，67：ll4～128．
Gross K L．A comparison of methods for estimating soil seed banks．Journal of Ecology，1990，78：1079—1093．
Poiani K A，Johnson W C．Evaluation of the emergence method in estima ting seed bank composition of prairie wetlands．Aqu atic Botany，1988，32：91～
97．
van der Valk A G，Pederson R L，Davis C B．Restoration and creation offreshwater wetlands using seed banks．Wetlands Ecology Manage，1992，l(4)：
191～ 197．
Brown S C．Remnant seed banks and vegetation as predictors of restored marsh vegetation．Canadian Journal of Botany， 1998，76：620—630．
Robe rts H A．Seed banks in soils．Advan ced Applied Biology ，1981，6：l一55．
Shannon C E，Weaver W ．Th e ma thema tical theory of Illinois：Illinois University Press，1949
Simpson E H ．Measurement of diversity．Nature，1949，163：688．
Hurlbe rt S H．Th e non-concept of species diversity：a critiqu e an d alternative parameters ．Ecology ，1971，52：577～586
Pielou E C．Ecology Diversity．New York：John Wiley＆ Sons Inc。1975．
Jaccard P．1The distribution oftheflora ofthe alpine zone．New Phytology ．1912．1l：37～5O．
Carter C T，Ungar I A． Aboveground vegetation，seed bank an d soil an alysis of a 3 1-year-old forest restoration on coal mine spo il in southeastern Ohio
．
American Midlan d Naturalist，2002，147：44～59．
Zhao L Y，Li F R，W an g X Z．Characteristics of soil seed bank an d standing vegetation chan ges in sandy grasslan d along a desertification gradient
． Acta
Ecologica Sinica，2003，23(9)：1745～1756．
Zhao L Y，Li F R，Zhang H，et 02 ．Characteristics ofthe soil seed bank at the fenced sandy meadow in Horqin Sandy Land．Chinese Journal of Ecology ，
2Oo4，23(2)：45～49．
O’Connor T G，Pickett G A．The influence of grazing on seed production an d seed bank of seme African savan na grasslan d
． Journal of Applied Ecology ．
1992，29：247—260．
Grime J P．Plant strategies and vegetation processes．Chichester：John Wiley and Sons，1979．
Leck M A，Graveline K J．1The seed bank of a freshwater tidal marsh．American ．Journal of Botony，1979，66：1006—1015．
Levassor C，Ortega M，Peco B．Seed banks dynamics of Mediteranean pastures subjected to mechanical disturbance．Journal of Vegetation Science，l99l，
l：339—344．
Maranon T．Soil seed bank and community dynamics in an annual-dominated Mediterranean salt-marsh．Journal ofVegetation Science，1998．9：371—378．
Henderson C B，Petersen K E，Redak R A．Spatial and temporal patterns in the seed bank and vegetation of a desert grassland community．Journal of
Ecology ，1988，76：717～728．
Peco B，Ortega M，Levasser C．Similarity between seed bank and vegetation in Mediterranean grassland：a predictive mode1．Journal of Vegetation Science，
1998，9：815～828．
Tang Y，Cao M，Zhang J H，et a1．Relation between soil seed bank and aboveground vegetation in tropical forest in Xishuangbanna． Chinese Journal of
Applied Ecology，1999，10(3)：279～282．
参考文献
6] 欧晓昆，金振洲．元谋干热河谷植被类型研究．云南植物研究，1987，87(3)：271 288
7] 束文圣，蓝崇钰，黄铭洪，等．采石场废弃地的早期植被与土壤种子库．生态学报，2003，23(7)：1305 1312．
21] 赵丽娅，李锋瑞，王先之．草地沙化过程地上植被与土壤种子库变化特征．生态学报，2003，23(9)：1745～1756．
22] 赵丽娅，李锋瑞，张华，等．科尔泌沙地围封砂质草甸土壤种子库特征的研究．生态学杂志，2004，23(2)：45～49
3O] 唐勇，曹敏，张建侯，等．西双版纳热带森林土壤种子库与地上植被的关系．应用生态学报 ，1999，10(3)：279 282
维普资讯 http://www.cqvip.com