RESEARCH COMMUNICATIONS

RESEARCH COMMUNICATIONS

Ecological consequences of rapid urban

expansion: Shanghai, China
Shuqing Zhao1, Liangjun Da2, Zhiyao Tang1, Hejun Fang2, Kun Song2, and Jingyun Fang1*
Since Chinas economic reform in the late 1970s, Shanghai, the countrys largest and most modern city, has
experienced rapid expansion and urbanization. Here, we explore its land-use and land-cover changes, focusing
on the impacts of the urbanization process on air and water quality, local climate, and biodiversity. Over the
past 30 years, Shanghais urban area and green land (eg urban parks, street trees, lawns) have increased dramatically, at the expense of cropland. Concentrations of major air pollutants (eg SO2, NOx, and total suspended particles) were higher in urban areas than in suburban and rural areas. Overall, however, concentrations have
decreased (with the exception of NOx), due primarily to a decline in coal consumption by industry and in private households. Increased NOx pollution was mainly attributed to the huge increase in the number of vehicles
on the roads. Water quality changes showed a pattern similar to that of air quality, with the most severe pollution occurring in urban areas. Differences in mean air temperatures between urban and rural areas also
increased, in line with the rapid pace of urban expansion, indicating an accelerating urban heat island effect.
Urban expansion also led to a decrease in native plant species. Despite its severe environmental problems,
Shanghai has also seen major economic development. Managing the tradeoffs between urbanization and environmental protection will be a major challenge for Chinese policy makers.

Front Ecol Environ 2006; 4(7): 341346

hina has been undergoing a period of economic reform

and expansion since the late 1970s, accompanied by
rapid and widespread urbanization. During the early 1980s,
18% of Chinas population lived in cities, but this rose to
39% by 2003, while the number of cities increased from 190
to 660 (including about 170 cities with a population greater
than 1 million) over the same time period. From 1980 to
2003, the contribution of Chinese cities to gross national
product (GNP) increased from 69.9% to 85.9% (Chinese
Statistical Bureau 2004). However, urbanization has also
led to serious environmental and ecological problems, both
in urban and surrounding areas, including increased air and
water pollution (Briant and Guo 2000; Liu and Diamond

Authors contact details are on p346)

The Ecological Society of America

2005; Shao et al. 2006), local climate alteration (Zhou et al.

2004), and a major reduction in natural vegetation cover
and production (Fang et al. 2003). Urban residents have
experienced an increase in levels of cholesterol-related diseases (Lee 2004) and an overall decline in quality of life.
As the largest and most modern city, Shanghai has
experienced extensive urban expansion over the past
three decades (Figure 1a). Between 1975 and 2003, the
citys population increased from 10.8 million to 13.4 million (Shanghai Statistic Bureau 2004). The resulting ecological consequences of urban sprawl have caused considerable concern among scientists and policy makers.
Several studies have documented some environmental
impacts of Shanghais increasing urbanization; for example, an analysis of meteorological data from both the citys
www.frontiersinecology.org

341

Urbanization in Shanghai

SQ Zhao et al.

of urbanization, we divided Shanghai

into three parts, according to the level
of urbanization: urban, suburban, and
rural (Figure 1b).

342

Remote sensing data and data

processing

Between 1975 and 2005, data on land

use and land cover in Shanghai was
obtained using cloud-free Landsat
Multispectral Scanner (MSS) and
Thematic Mapper (TM) remote sensing images. The data, which spanned
three decades, covered seven periods:
1975, 1981, 1987, 1990, 1995, 2000,
and 2005 (mostly from April to June).
The first two sets of satellite images
(1975 and 1981) were obtained from
the MSS; for all other years, the TM
Figure 1. The city of Shanghai. (a) An aerial view of the downtown area. (b and inset) was used. Landsat data were interMap showing the location of Shanghai, at the estuary of the Yangtze River, on the coast of preted using ERDAS 8.4 image-prothe East China Sea. For the purposes of the current study, Shanghai was divided into three cessing software. Land-cover data were
areas: (A) urban; (B) suburban; and (C) rural.
divided into four types: urban land
(urbanized area), green land (ie urban
urban and adjacent rural areas has revealed the presence parks, street trees, lawns, etc), cropland, and water body. For
of a heat island effect (higher air temperatures in urban details on the data processing, see Fang et al. (2005).
areas than in the surrounding areas; Chow 1992; Chen et
al. 2003). Water quality in the citys central business dis- Air and water quality data
trict has been seriously degraded (Ren et al. 2003), and soil
in the Baoshan District is often contaminated with lead, Systematic monitoring data on air and water pollution in
zinc, and cadmium (Hu et al. 2004). Ye et al. (2000) have Shanghai are available from 1983 onwards (Shanghai
suggested that the huge increase in the number of motor Environmental Protection Bureau 19832004). To meavehicles associated with the urbanization process has led sure air quality, data were collected on sulfur dioxide
(SO2), acid rain frequency, total suspended particles
to serious human health risks in the city.
To date, however, long-term (spanning both pre- and (TSP), and nitrous oxides (NOx); for water quality, the
post-urbanization periods) and spatially explicit monitor- data covered dissolved oxygen (DO), chemical oxygen
ing of the urbanization of the entire Shanghai area, demand (COD), biochemical oxygen demand (BOD),
together with comprehensive studies of the ecological Kjeldahl nitrogen (KN), total phosphorus (TP), oils,
consequences, have not been conducted. Here, we docu- volatile phenol (VP), and total mercury (THg). We anament the processes of urbanization in Shanghai between lyzed variations in the concentrations of these components
1975 and 2005, through the analyses of satellite-derived in urban, suburban, and rural areas between 1983 and
data regarding land-use and land-cover changes. We 2004. With the exception of DO, all these measures of air
explore the changes in air and water quality, the urban and water quality are inverse indices; that is, the higher the
heat island effect, and the biodiversity loss, and examine values, the worse the air or water quality.
their possible associations with urban expansion.
Climatic data
 Data and methods
Urbanization leads to alterations of the local climate, and
in particular creates a significant heat island effect
Site description
(Kalnay and Cai 2003; Zhou et al. 2004). We collected
Shanghai is located on the coast of the East China Sea, at the monthly mean air temperatures at 10 meteorological staestuary of the Yangtze River. The climate is subtropical, with tions for the period between January 1975 and December
an annual precipitation of 1200 mm and a mean annual tem- 2004. Four of these stations were located in urban areas
perature of 16 C. The total area covered by the city is 8010 (Xujiahui [Shanghai proper], Minhang, Jiading, and
km2 (latitude: 30 40 to 31 55 N; longitude: 120 50 to Baoshan), with the remainder in rural areas (Nanhui,
121 55 E). In order to examine the ecological consequences Fengxian, Jinshan, Qingpu, Chongming, and Songjiang;
www.frontiersinecology.org

The Ecological Society of America

SQ Zhao et al.

Urbanization in Shanghai
(a) 1975

Biodiversity data

(d) 1990

Urbanization exerts a substantial effect on biodiversity, resulting in the loss of native species and the introduction of nonnative species (Blair 1999; McKinney 2000). To explore the
impact of urbanization on biodiversity, we collected species
richness data for flora and fauna from various literature
sources (Xu et al. 1999; Yang et al. 2002; Shanghai Agriculture
and Forestry Bureau 2004). We also carried out fieldwork to
investigate the numbers of native and non-native woody
plant species in different green areas (Da et al. 2005).

(b) 1981

(c) 1987

(e) 1995

(f) 2000

(h) Urban expansion 19752005

Legends

(g) 2005
Legends

1975
19751981
19811987
19871990
19901995
19952000
20002005

Water body
Cropland
Green land
Urban land

 Results and discussion

Land-use and land-cover change

Air and water quality changes

Urbanization places a heavy burden on local air and water

quality. Air quality monitoring data obtained from different
observatories showed that concentrations of SO2, TSP, and
NOx, along with acid rain frequency, varied considerably in
different areas along the urbanization gradient.
Concentrations of air pollutants were consistently highest
in urban areas, lower in suburban areas, and lowest in rural
areas over the study period, indicating a strongly negative
The Ecological Society of America

(i)
Area of urban land, green
land, and water body (km2)

Land use and land cover in Shanghai have been greatly

altered over the past three decades, as a result of the rapid
expansion of urban areas (Figure 2). The area of urban
land increased from 159.1 km2 in 1975 to 1179.3 km2 in
2005 (Figure 2i). There were distinct phases in the urbanization process. The slowest rate of urbanization, 17.7 km2
yr1, occurred between 1975 and 1981 (Figures 2a and
2b), increasing to 52.4 km2 yr1 between 1990 and 1995
and 54.9 km2 yr1 from 2000 to 2005 (Figures 2d to 2g).
This is consistent with Chinas economic policies, since
the country began its economic reform in 1978, and
accelerated the process in 1992 (Lin 1999).
Urbanization is generally associated with an increase in
managed green areas, such as street trees, lawns, and parks
for urban recreation; these improve both the visual appeal
of a city and environmental quality (Attwell 2000). In parallel with its urban expansion, Shanghais green areas have
continued to increase in size, from 8.7 km2 in 1975 to 252.9
km2 in 2005. In contrast, agricultural land area has fallen
rapidly, from 6030.7 km2 in 1975 to 4743.1 km2 in 2005. In
addition, the area covered by water has shown small fluctuations over the past three decades (Figure 2i).

2500

Urban land
Water body

Green land
Cropland

6500
5500

2000

4500

1500

3500
1000

2500

500

1500

Area of cropland (km2)

Figure 1b). We analyzed the differences in mean annual

temperature (MAT), monthly mean maximum temperature (MTmax), and monthly mean minimum temperature
(MTmin) between the urban and rural areas. In order to
study the urban heat island effect, we correlated the area
of urban land with the differences in mean temperatures
between the urban and rural areas from 1975 to 2005.

500
1975 1981 1987 1990 1995 2000 2005
Year

Figure 2. Changes in land-cover types in Shanghai. (ag)

Spatial distribution of land-cover types for seven time periods:
1975, 1981, 1987, 1990, 1995, 2000, and 2005. (h) Overlap
of urban lands from 19752005. (i) The changes in area of four
land-cover types: urban land, green land, cropland, and water,
from 19752005.

influence of major urban development on air quality

(Figure 3). The temporal variations in air pollutants suggest
that SO2, TSP, and acid rain were largely the result of coal
combustion, since levels of these pollutants decreased from
1983 to 2004 in all three areas, while NOx concentration
increased (Figure 3). The falling concentrations were
mainly related to the decline in the use of coal, both by
industry and in private households (Yuan and James 2002).
The increase in green lands may also have helped to mitigate air pollution. The rising concentrations of NOx are
attributed primarily to a rapidly increasing number of motor
vehicles in recent years; for example, vehicles in Shanghai
numbered less than 100 000 in the early 1980s, but
rose to more than 2 million by 2004 (www.shtaq.com/
www.frontiersinecology.org

343

Urbanization in Shanghai
Acid rain frequency (%)

improvement program, aimed at moving

industrial units from urban centers to
suburban areas in order to improve urban
air quality (Shanghai Environmental
Protection Bureau 2001).
The second exception is that NOx pollution has decreased in urban areas, since
(c)
(d)
the late 1990s, a result of the Clean
Vehicle Program which was implemented in 1999 as a means of controlling
air pollution due to vehicle emissions
(He et al. 2002).
Water pollution levels show similar
patterns, being consistently more severe
in the central urban areas than in suburFigure 3. Changes in major components of air quality between 1983 and 2004 in ban and rural areas (Figure 4). During the
the different areas of Shanghai: (a) sulfur dioxide (SO2); (b) acid rain frequency; study period, different pollutants varied
(c) total suspended particles (TSP); and (d) nitrous oxides (NOx).
over time in the central urban areas; concentrations of COD, BOD, KN, and
showcontent/contentdetail_old.asp?id=39703). A similar volatile phenols increased from the early 1980s to the
result was seen in Beijing (He et al. 2002).
early 1990s, and then began to decrease, while DO, total
There have been two exceptions to these general trends. phosphorus, oils, and total mercury did not exhibit signifiFirst, as shown in Figures 3a and 3b, SO2 pollution and cant changes over the same period. This suggests that
acid rain have tended to increase in the suburbs in recent water quality in the city center has been improving since
years, which may be attributed to a growing number of fac- the 1990s. However, water quality in the suburban and
tories in suburban areas. In the late 1990s, the municipal rural areas has deteriorated over the past decade, although
Government of Shanghai initiated an environmental generally much better than in the central urban area. For
example, BOD, KN, and TP pollution has increased in
recent years, probably as a result of the transfer of factories
(b)
(a)
from the city center to suburban and rural areas (Shanghai
Environmental Protection Bureau 2001).
(b)

DO (mg L1 )

COD (mg L1 )

NOx (mg m3 )

TSP (mg m3 )

SO2 (mg m3 )

(a)

(e)

(g)

KN (mg L1 )
Oils (mg L1 )

(c)

THg (103mg L1 )

TP (mg L1 )

BOD (mg L1 )

Local climate change and the heat island effect

VP (mg L1 )

344

SQ Zhao et al.

(d)

(f)

(h)

Figure 4. Changes in major components of water quality

between 1983 and 2004 in the different areas of Shanghai: (a)
dissolved oxygen (DO); (b) chemical oxygen demand (COD);
(c) biochemical oxygen demand (BOD); (d) Kjeldahl nitrogen
(KN); (e) total phosphorus (TP); (f) oils; (g) volatile phenols
(VP); and (h) total mercury (THg).
www.frontiersinecology.org

As illustrated in Figure 5, a substantial urban heat island

effect was found in Shanghai. The difference in MAT
between urban and rural areas increased from 0.1 C in
the late 1970s (average for 19751979; a mean MAT of
15.7C vs 15.6C at urban and rural stations) to 0.7C in
the early 2000s (average for 20002004; 17.3C vs
16.6C), with an increase of 0.24C per decade (Figure
5a). Similarly, MTmax and MTmin did not show a difference between urban and rural areas in the late 1970s
(28.2C vs 3.4C for MTmax and MTmin, respectively),
although the differences increased to 0.7C (29.2C vs
28.5C for the urban and rural area) and 0.5C (5.0C vs
4.5C) in the early 2000s, respectively, with a decadal
increase of 0.26C and 0.21C (Figures 5b and 5c).
The correlation analysis of the relationship between
the differences in mean temperatures in urban vs rural
areas, and the amount of urban land, indicate that the
differences in temperature between urban and rural areas
increased substantially as the city expanded, and that this
increase was faster for MTmax than for MTmin (Figure 5d).
This strongly suggests that rapid urbanization can
increase temperatures considerably in the city and adjacent areas (Zhou et al. 2004).
The Ecological Society of America

SQ Zhao et al.

Urbanization in Shanghai

Changes in biodiversity

345

(b)

(a)

The area around Shanghai is rich in

biodiversity. There are 1904 spermatophyte plants, representing 981 genera
and 168 families. However, there are
more non-native species than native
(968 and 936, respectively; Xu et al.
1999). Seven hundred and sixty vertebrate species are found in Shanghai, of
(c)
(d)
which 40 are mammals, 424 birds, 32
reptiles, 14 amphibians, and 250 fish
(Shanghai Agriculture and Forestry
Bureau 2004).
Urban expansion and the associated
increase in human activities have led to
a considerable loss of biodiversity in the
study area. The number of native plant
species has fallen rapidly in the relatively wild regions; for example, the Figure 5. Changes in temperature differences between urbanized and rural areas in
number of plant species in the Sheshan the years 19752004. (a) Difference in mean annual temperature (dMAT); (b)
area fell from 535 in the 1980s to 254 by difference in monthly mean maximum temperature (dMTmax); (c) difference in
the end of the 1990s (Xu et al. 1999), monthly mean minimum temperature (dMTmin); and (d) relationships between the
while on Dajinshan Island plant species temperature differences and degree of urbanization.
declined from 254 in the 1980s to 145
in 2000 (Yang et al. 2002). At the same time, the number gether with an improvement in public awareness of enviof non-native plant species has increased greatly due to ronmental issues, has been the basis for attempts to achieve
the introduction of many species into the managed green both socioeconomic and environmental sustainability in
spaces; about 300 alien plant species were introduced to the urbanization process. In recent years, the Shanghai
the Shanghai area between 1980 and 2005, for instance. Government has implemented a series of measures specifiIn 2004, a comprehensive investigation by the authors cally designed to protect the local environment. First, a
recorded a total of 206 woody plant species within the number of policies have been developed which are aimed at
green spaces in the city, of which only 64 (31.1%) were promoting clean energy use to reduce water and air pollunative to Shanghai and adjacent areas (Da et al. 2005). tion; over the past 10 years, these have involved improveThe proportion of native species is even lower in recently ment of the sewage treatment infrastructure, removal of
established green areas within the city. For example, in exhaust emission sources, improvement of transportation
the newly created Yanzhong green land, only 26 (18.4%) systems, and control of vehicle density. As a result, air and
of 142 woody plants are native, while a much higher pro- water quality have begun to improve, beginning in the
portion (28 out of 69 species) of woody plant species were 1990s (Figures 3 and 4). Second, afforestation and the
found in the green spaces established in the 1950s on the establishment of parks within Shanghai have been incorporated into city planning, leading to an increase in vegetacampus of East China Normal University (Table 1).
tion coverage. For example, the area occupied by green
land in Shanghai has risen from 870 ha in 1975 to 25 300 ha
 Conclusions
Shanghai has experienced
rapid urbanization over the
past three decades, accompanied by large-scale economic development. However, this urban growth has
caused a number of ecological problems, including
degradation of air and
water quality, alteration of
the local climate, a decline
in native species, and an
increase in numbers of
alien species. This, to The Ecological Society of America

Table 1. Comparison of the number and proportion of native woody plant species in
three different green lands in Shanghai
Site

No. of
woody plants

No. (%) of species

native to Shanghai

No. (%) of species native

to the adjacent areas

No. (%) of
native species

All green lands in the

urban center

206

43 (20.9%)

21 (10.2%)

64 (31.1%)

Newly established
Yanzhong green land

142

11 (7.8%)

15 (10.6%)

26 (18.4%)

69

16 (23.2%)

12 (17.4%)

28 (40.6%)

Green land on the East

China Normal University
campus, established in
the 1950s
Based on Da et al. (2005)

www.frontiersinecology.org

Urbanization in Shanghai

346

in 2005 (Figure 2). Third, nature reserves and forested parks

have been established to protect biodiversity and provide
recreation sites. The first nature reserve in Shanghai was
established in the early 1990s, followed by four more reserves
and a national forest park. Finally, in 2003, Chongming
Island, the third largest island in China and the largest alluvial island in the world, was declared an ecological island
by the Shanghai Government. Chongming Island, situated
at the estuary of the Yangtze River in northern Shanghai,
covers an area of 1100 km2 and includes a large coastal wetland and tidal flats that provide habitat for a wide variety of
species. All development activities that conflict with the
protection of the environment will be prohibited on the
island. At the same time, the Shanghai Government provides monetary compensation to local communities and
encourages local resident participation in the protection of
the island (Yuan et al. 2003).
In summary, China is facing many challenges as its
urban growth rate continues to accelerate. Rapid urbanization has greatly accelerated economic and social development, but has also created severe environmental problems. Maintaining a balance between environmental
sustainability and the continuing process of urbanization
is a major issue facing the Chinese Government. A
national strategy for sustainable development Chinas
Agenda 21 (Department of Planning Committee of China
1994) has been established, aimed at reducing the negative environmental impacts of economic development
while at the same time maintaining economic and social
benefits. This involves land-use management strategies
and the development of greener cities, thereby providing a
healthier environment for both humans and wildlife.

 Acknowledgment
We thank LN Li, NN Liu, XY Wang, JQ Qian, L Li, JY
Wang, CX Wang, and KK Shang from East China Normal
University for their help in data collection. This study was
funded by the State Key Basic Research and Development
Plan of China (#G2000046801), the National Natural
Science Foundation of China (#40571150 and
#40024101), and Special Research Fund for the Doctoral
Programme of Higher Education (#20050269015).

 References

Attwell K. 2000. Urban land resources and urban planting case

studies from Denmark. Landscape Urban Plan 52: 14563.
Blair RB. 1999. Birds and butterflies along an urban gradient: surrogate taxa for assessing biodiversity? Ecol Appl 9: 16470.
Briant LD and Guo J. 2000. Airborne particulate study in five cities
of China. Atmos Environ 34: 270311.
Chen LX, Zhu WQ, Zhou XJ, and Zhou ZJ. 2003. Characteristics of
the heat island effect in Shanghai and its possible mechanism.
Adv Atmos Sci 20: 9911001.
Chinese Statistical Bureau. 2004. Chinese statistical year book.
Beijing, China: Chinese Statistical Press.
Chow SD. 1992. The urban climate of Shanghai. Atmos Environ B
Urb 26: 915.
Da LJ, Fang HJ, Chen KX, et al. 2005. Woody plant species in green
www.frontiersinecology.org

SQ Zhao et al.
lands in Shanghai. Special report on investigation of woody
plant species in the urban area in Shanghai. Shanghai, China:
East China Normal University.
Department of Planning Committee of China. 1994. Chinas
Agenda 21 white paper on Chinas population, environment,
and development in the 21st century. Beijing, China:
Environmental Science Press.
Fang JY, Piao SL, Field CB, et al. 2003. Increasing net primary production in China from 1982 to 1999. Front Ecol Environ 1:
29397.
Fang JY, Rao S, and Zhao SQ. 2005. Human-induced long-term
changes in the lakes of the Jianghan Plain, Central Yangtze.
Front Ecol Environ 3: 18692.
He KB, Huo H, and Zhang Q. 2002. Urban air pollution in China:
current status, characteristics, and progress. Annu Rev Energ
Env 27: 397431.
Hu XF, Wu HX, Hu X, et al. 2004. Impact of urbanization on
Shanghais soil environmental quality. Pedosphere 14: 15158.
Kalnay E and Cai M. 2003. Impact of urbanization and land-use
change on climate. Nature 423: 52831.
Lee LM. 2004. The current state of public health in China. Annu
Rev Publ Health 25: 32739.
Lin GCS. 1999. State policy and spatial restructuring in postreform China, 197895. Int J Urban Regional 23: 67096.
Liu JG and Diamond J. 2005. Chinas environment in a globalizing
world. Nature 435: 117986.
McKinney ML. 2002. Urbanization, biodiversity, and conservation. BioScience 52: 88390.
Ren WW, Zhong Y, Meligrana J, et al. 2003. Urbanization, land
use, and water quality in Shanghai, 19471996. Environ Int 29:
64959.
Shanghai Agriculture and Forestry Bureau. 2004. Terrestrial wild
plants and animals resources in Shanghai. Shanghai, China:
Shanghai Scientific and Technology Press.
Shanghai Environmental Protection Bureau. 19832004. Report
on environmental quality of Shanghai. Shanghai, China:
Shanghai Environmental Protection Bureau.
Shanghai Environmental Protection Bureau. 2001. Report on
environmental quality of Shanghai (19962000). Shanghai,
China: Shanghai Environmental Protection Bureau.
Shanghai Statistical Bureau. 2004. Shanghai statistical yearbook.
Shanghai, China: Chinese Statistical Press.
Shao M, Tang XY, Zhang YH, et al. 2006. Environmental pollution
of city clusters in China: current situation and challenges.
Front Ecol Environ 4: 35361.
Xu BS, Ou SH, and Yang BS. 1999. Flora of Shanghai. Shanghai,
China: Shanghai Scientific and Technology and Document Press.
Yang YC, Da LJ, and Qin XK. 2002. A study of the flora on
Dajinshan Island in Shanghai. Chinese J Wuhan Bot Res 20:
43337.
Ye SH, Zhou W, Song J, et al. 2000. Toxicity and health effects of
vehicle emissions in Shanghai. Atmos Environ 34: 41929.
Yuan W and James P. 2002. Evolution of the Shanghai City region
19781998: an analysis of indicators. J Environ Manage 64:
299309.
Yuan W, James P, Hodgson K, et al. 2003. Development of sustainability indicators by communities in China: a case study of
Chongming County, Shanghai. J Environ Manage 68: 25361.
Zhou LM, Dickinson RE, Tian YH, et al. 2004. Evidence for a significant urbanization effect on climate in China. P Natl Acad
Sci USA 101: 954044.
1

Department of Ecology, College of Environmental Sciences, and Key

Laboratory for Earth Surface Processes of the Ministry of Education,
Peking University, Beijing, 100871, China *(jyfang@urban.pku.edu.cn);
2
Department of Environmental Sciences, and Shanghai Key Laboratory for
Ecology of Urbanization Process and Ecorestoration, East China Normal
University, Shanghai, 200062, China
The Ecological Society of America

RESEARCH COMMUNICATIONS

RESEARCH COMMUNICATIONS
347

Biodiversity in Chinas mountains

Zhiyao Tang, Zhiheng Wang, Chengyang Zheng, and Jingyun Fang*
China, one of the worlds megabiodiversity countries, is home to more than 30 000 vascular plant and 6300
vertebrate species. Over thousands of years, however, cultivation has led to the disappearance of many of these
species from the plains and lowland areas. The mountain regions still harbor large numbers of species, because
there have been fewer human and natural disturbances and there are more diverse habitats. We used county
level species distribution databases to explore patterns of biodiversity and to identify biodiversity hotspots
within China. Ten hotspot ecoregions were identified, containing 3110 plant genera (92.0% of the countrys
total), 220 (90.5%) endemic plant genera, 366 (94.3%) endangered plants, and 254 (72.2%) endangered vertebrates, 427 (91.0%) terrestrial mammal species, and 65 (85.5%) endemic mammals. All 10 hotspot ecoregions
are located in the mountainous areas of China. Although high richness of overall, endangered, and endemic
plants and animals co-occurred in many of the same hotspot ecoregions, they often occurred in different counties within these ecoregions and showed low spatial congruence. In conclusion, Chinas mountain regions are
critical for protecting biodiversity and should be made conservation priorities in the future.

Front Ecol Environ 2006; 4(7): 347352

hina spans a huge geographical area, ranging from tropical to boreal zones and from very low altitudes (156 m
below sea level in the Turpan Desert, Xinjiang) to the
worlds highest mountain, Mount Qomolangma (Everest),
on the border of Tibet and Nepal. Almost all of the different types of biomes found on Earth, from rainforests to
deserts, are found in China. This geographic diversity provides abundant habitats for plants and animals. Home to
more than 30 000 vascular plants (surpassed only by Brazil
and Colombia) and 6300 vertebrates, China is one of the
worlds megabiodiversity countries (McNeely et al. 1990).
However, this biodiversity has suffered severe degradation
due to the density of the human population, a long history
of cultivation, and an increase in the intensity and extent of
human disturbances in the plains and lowland areas.
The mountain regions of China provide rich habitats
for much of the countrys remaining biodiversity, owing
to the heterogeneity of climates and soils, rapid elevational changes, varying aspects of slope direction, abundant microhabitats, and limited suitability for cultivation

The Ecological Society of America

(Krner and Spehn 2002). Thus, Chinas mountains are

likely to be especially important for preserving its remaining biodiversity (Chen 1998).
In this study, we used well-developed county level distribution data for plants and animals to explore the patterns of richness of plant genera and terrestrial mammals,
endemic plant genera and mammal species, and endangered plant and vertebrate species in China. Our goals
were to identify hotspots for three different aspects of biodiversity (overall, endemic, and endangered) and to
investigate the potential role of the mountain regions in
conserving Chinas biodiversity.

 Data sources
There are 2383 counties in China. A county level map
was digitized and overlaid with a Digital Elevation Model
(DEM) with 1 km x 1 km resolution, obtained from the
United States Geological Survey (USGS), to document
the topographic attributes of each county.
www.frontiersinecology.org

Biodiversity in Chinas mountains

348

The overall numbers of plant genera and mammal

species, endemic plant genera and mammal species, and
endangered plant and vertebrate species for each county
were derived from the following resources: Seed plants of
China (Wu and Ding 1999), Mammal species in
China (Zhang 1997), China red data book of endangered plants (Fu 1992) and China red data book of
endangered animals (Wang 1998).
The Seed plants of China database (Wu and Ding
1999) is the main product of a national project spanning
10 years (19901999) that described county-level distributions for each of the 32 308 native and cultivated seed
plants, representing 363 families and 3428 genera (including 243 genera endemic to China). For each genus, the
geographic distribution was resolved to the county level,
and habitat type, elevational range, and longitudinal and
latitudinal range were documented based on national
flora, local reports, and other scientific literature. Because
this database is derived primarily from plant specimen
records, it is potentially biased by regional differences in
levels of investigation. In the study presented here, genera
(or family) richness was used instead of species richness to
measure plant diversity, because it is less likely that these
higher taxa would be overlooked in surveys (Qian 1998).
To help reduce the possibility of bias in the records as a
result of regional differences in investigation intensity, we
explored the patterns of seed plant density at the genus, as
opposed to species, level in this study.
Mammal species in China (Zhang 1997) was compiled
based on the scientific literature as well as observations of
local fauna over the past 50 years. This book includes information on the county level distribution and habitat type for
each of the 537 mammals in China. In our study, 68 marine
mammals whose distributions are not clear were excluded
from the analysis. We therefore included only the data for
469 terrestrial mammal species (of which 77 are endemic to
China) to investigate patterns of mammal diversity.
The China red data book of endangered plants (Fu
1992) and the China red data book of endangered animals (Wang 1998) provided information on the distributions of endangered plant and vertebrate species.
These two datasets record the geographic distribution,
habitat type, and elevational range for each of Chinas
388 endangered plant species and 352 endangered vertebrate species (including 134 mammals, 95 reptiles, 92
fishes, and 31 amphibians). Endangered birds were also
not included in this study as most avian species are
migratory, thus making their distributions unclear.

 Methods
We first calculated the animal and plant richness per county
for the three aspects of biodiversity mentioned above (overall, endemic, and endangered). Geographic information system (GIS) software (ArcView GIS 3.2) was used to calculate
the area covered by each county. The administrative division atlas of China (Editorial Committee for the
www.frontiersinecology.org

ZY Tang et al.

Administrative Division Atlas of China 2002) was digitized

to obtain information on county boundaries. The area covered by the different counties in China varies greatly; we
used the following area-adjusted species (genera for overall
and endemic plants) density as a measure of diversity, to
eliminate the influence of county size (Qian 1998):
D = S/ln (A)
where D is the species density for a county, S is the number of species, and A is the area of the county.
We then used ArcView GIS 3.2 to map different aspects
of biodiversity in order to identify biodiversity hotspots
across the country. Here, we defined hotspot counties as
the top 5% of land area with the highest density of each of
the three aspects of biodiversity. We first selected the
county that contained the highest density of species, then
added the county with the second highest density, then
the third, and so on, until the total area of selected counties equaled 5% of Chinas total land area. All of the counties ranked within the top 5% for each aspect of biodiversity were designated as the hotspot counties. Because
ecoregions can be used as conservation units at regional
scales (Olson and Dinerstein 1998), we assigned all the
counties to different ecoregions, according to their physical geography (Editorial Committee for Chinas Physical
Geography 1985). We then calculated the percentage of
the area covered by hotspot counties within each ecoregion. An ecoregion was designated as a hotspot if 5% of its
total area was covered by hotspot counties.
The topographic attributes (maximum and minimum
elevation, average elevation, elevational range, and
slope) of each county were derived from the DEM by
using ArcView GIS 3.2. The UNEP-WCMC (2002) criterion was used to identify mountainous areas versus
plains. According to the criterion, areas that fit any of the
following four descriptions are defined as mountainous:
(1) an elevation of 3001000 m and a local elevational
range of > 300 m; (2) an elevation of 10001500 m above
sea level (asl), and a slope of > 5 or a local elevational
range of > 300 m asl; (3) an elevation of 15002500 m asl,
and a slope of > 2; or (4) an elevation of > 2500 m asl.

 Results and discussion

Macrotopography in China

According to the UNEP-WCMC (2002) definition,

mountains cover a total area of approximately 4.6 million
km2 in China (48% of the total land area). Figure 1 shows
the macrotopography and a frequency distribution of the
mean elevation per county, together with major geographic regions. Low-lands with an elevation of < 500 m
asl covered 2.63 million km2 (27.4% of the total land
area), while areas with an altitude of > 1000 m asl encompassed an area of 5.42 million km2 (56.5% of the total
land area). Another 1.93 million km2 (20.1% of the
The Ecological Society of America

ZY Tang et al.

Biodiversity in Chinas mountains

countrys total area) was covered by altitudes of > 4000 m asl.

349

Patterns of overall plant genera and

mammal species densities

>60

5560

5055

4550

4045

3540

3035

2530

2025

1520

1015

>5

The Ecological Society of America

510

Area (x106 km2)

The density of plant genera was higher in

eastern compared to western China, higher
in southern compared to northern China,
and also in the mountainous areas comElevation (m)
pared to the plains and plateaus (Figure
<200
200500
2a). One hundred and seventy-three coun5001000
10002000
ties were assigned as hotspot counties for
20003000
3
30004000
plant genera; they contained densities of
40005000
50006000
> 65 genera ln1 (km2). In total, 2948 gen2
<6000
era of seed plants (86.0% of Chinas total)
1
grow in these counties. The counties were
located mostly within eight regions: the
0
Hengduan Mountains, the Xishuangbanna
area, the upper reaches of the Hongshui
Elevation (x100 m)
River (the Wumeng Mountains), the
Nanling Mountains, the Wuling
Mountains, the East China Mountains, Figure 1. Macrotopographic pattern in China, together with locations of major
Hainan Island, and Taiwan Island. mountains. The inset figure illustrates the elevationarea frequency distribution.
Compared to the mountainous areas listed
above, counties in the North China Plain, Loess Plateau, Patterns of endemic plant genera and mammal
Sichuan Basin, and Qinghai-Xizang Plateau contained species densities
The density of endemic plant genera was also higher in
very low densities of plant genera (< 20 genera ln1 [km2]).
Mammal species densities showed a slightly different pat- mountainous areas than in other areas. Altogether, 161
tern from that of plant genera. One hundred and fifty-four counties were assigned as hotspots for endemic plants
counties were designated as hotspot counties, with mam- (> 0.85 genera ln1 [km2]). These counties contained 204
mal species densities of > 5.8 species ln1 (km2). These endemic plant genera (84.0% of the total), with most
counties contained 384 mammals (81.9% of Chinas total) located in the following seven regions: the Hengduan
and were mostly located in the following eight regions: the Mountains, the Wuling Mountains, the Nanling
Hengduan Mountains, the Nanling Mountains, the Mountains, the Qinling Mountains, the East China
Xishuangbanna Area, the Qinling Mountains, the upper Mountains, Hainan Island, and Taiwan Island (Figure 2c).
reaches of the Hongshui River, the East China Mountains, On the other hand, the 123 hotspot counties for endemic
Hainan Island, and Taiwan Island. Densities of mammal mammals (a density of > 0.75 species ln1 [km2]) contained
species were lower in the North China Plain, the Central 56 endemic mammals (73.7% of the total), which were
Yangtze Plain, the Loess Plateau, the western part of the located primarily in the Hengduan Mountains, the Qinling
Qinhai-Xizang Plateau, the Northeast China Plain, and Mountains, and the eastern part of the Qinghai-Xizang
Plateau (Figure 2d). These mammal hotspot counties were
the Sichuan Basin (Figure 2b).
The distribution of overall richness is positively associ- located either in the ecotones between different biomes or
ated with habitat diversity (Orme et al. 2005). In the pre- in areas with high environmental heterogeneity, and have
sent study, we used the elevation range (ie the difference not suffered the influences of the last period of glaciation
between the highest and the lowest
elevation in a county) as a surrogate of Table 1. Correlation between densities of overall and endemic plant genera,
habitat heterogeneity, and found posi- overall and endemic mammal species, and endangered plant and vertebrate
tive relationships between elevation species in different counties of China
range and the densities of both plant
Endemic
Endangered Overall
Endemic
Endangered
plant genera plants
mammals mammals
vertebrates
genera (Pearson correlation coeffi- Aspect of diversity
cient, r = 0.32, P < 0.01), and mammal Overall plant genera 0.41*
0.42*
0.36*
0.07
0.41*
0.44*
0.31*
0.39*
0.34*
species (r = 0.36, P < 0.01). Statistical Endemic plant genera
0.38*
0.26
0.45*
analysis also suggested a positive rela- Endangered plants
0.57*
0.54*
tionship between the density of over- Overall mammals
Endemic mammals
0.31*
all plant genera and mammal species
*P < 0.01. For all correlations, n = 2383 counties.
(r = 0.36, P < 0.01; Table 1).
www.frontiersinecology.org

Biodiversity in Chinas mountains

350

ZY Tang et al.

ing birds) are similar to that of endangered plants. For these species, 143 counties were assigned as hotspot counties
(> 2.2 species ln1 [km2]); they contained
225 endangered vertebrates (63.9% of the
total). These counties were located in the
following eight regions: the Hengduan
Mountains, upper reaches of the Honghe
Plants (ln [km ])
Mammals (ln [km ])
020
01
River area, the Xishuangbanna area, the
2040
13
4065
35.8
Qinling Mountains, the Nanling
65100
5.88
>100
>8
Mountains, the East China Mountains,
Hainan Island, and the east part of the
(c)
(d)
Qinghai-Xizang Plateau. Low densities of
endangered vertebrates were found in the
northern and northeast plains and
plateaus in China (Figure 2f).
Endangered species richness is widely
used as a proxy for conservation priority
Endemic mammals (ln [km ])
Endemic plants (ln [km ])
identification because of the significant
00.2
00.2
0.20.4
0.20.5
and positive relationship between endan0.40.75
0.50.85
0.751.25
0.852
>1.25
>2
gered species and overall species richness
(Dobson et al. 1997). As in previous studies (eg Dobson et al. 1997), we observed a
(e)
(f)
positive correlation between densities of
overall mammals and endangered vertebrates (r = 0.54, P < 0.01), and between
densities of overall plant genera and
endangered plants (r = 0.42, P < 0.01;
Table 1). Further, positive correlations
Endangered plants (ln [km ])
Endangered vertebrates (ln [km ])
were found between the densities of
01
00.5
11.5
0.51
1.52.05
12.2
endangered plants and endemic plant
2.053
2.24.5
>3
>4.5
genera (r = 0.44, P < 0.01) and between
endangered and endemic animals (r =
Figure 2. Distribution patterns for the county level density of: (a) seed plant genera; (b) 0.31, P < 0.05; Table 1), as most of the
mammal species; (c) endemic plant genera; (d) endemic mammal species; (e) endangered endemic plants and vertebrates in China
plant species; and (f) endangered vertebrate species (excluding birds) in China. Hotspot are endangered because of their limited
counties are shown in dark orange and brown.
geographic distributions (Wang 1998).
We also found that the densities of
(eg the Hengduan Mountains, the Qinling Mountains) or endangered plant and vertebrate species were positively
are areas with high rates of speciation (eg the Tibetan correlated (r = 0.45, P < 0.01; Table 1).
Plateau; Chen 1998).
Richness of endangered species is shaped by the interaction between the biological mechanisms promoting
species diversity and the anthropogenic mechanisms
Patterns of endangered plant and vertebrate
eroding that diversity (Orme et al. 2005). To explore the
densities
influence of human threat on the distribution of endanHotspots for endangered plants included 185 counties gered species richness, we calculated the ratio of endan(> 1.1 species ln1 [km2]). These counties contained a total gered compared to overall densities (endangered plants to
of 333 endangered plants (85.8% of the total) and were overall plant genera and endangered vertebrates to overlocated in the following five regions: the Hengduan all mammals). This ratio was high in the North China
Mountains, the Xishuangbanna Area, upper reaches of the Plain for plants, and in the East China Mountains, the
Hongshui River area, Hainan Island, and Taiwan Island. Wuling Mountains, and the Nanling Mountains for aniCompared with these mountainous areas, the counties in mals. The North China Plain, the east (including the
the North China Plain, Northeast China Plain, Central East China Mountains), the central (where the Wuling
Yangtze Plain, Sichuan Basin, Loess Plateau, Inner Mountains are located), and the southern (including the
Mongolia Plateau, and Qinghai-Xizang Plateau possess Nanling Mountain) regions of China have high human
very low densities of endangered plant species (Figure 2e).
population densities, suggesting the influence of humans
The density patterns of endangered vertebrates (exclud- on species endangerment (Dobson et al. 1997).
(b)

(a)

www.frontiersinecology.org

The Ecological Society of America

ZY Tang et al.

Biodiversity in Chinas mountains

gered plant and vertebrate species appeared

in different counties, which were aggregated
in several ecoregions (Editorial Committee
215 497
200 328
for Chinas Physical Geography 1985). These
(22.0%)
(20.0%)
encompassed the Hengduan Mountains, the
80 832
111 507
59 037
77 879
Wuling Mountains, the Nanling Mountains,
(8.3%)
(11.1%)
(6.0%)
(7.8%)
upper reaches of the Hongshui River area, the
82 080
112 281
(11.5%)
(8.2%)
211 846
231 801
220 797
254 852
Xishuangbanna area, the Qinling Mountains,
(21.7%)
(23.7%)
(22.0%)
(25.5%)
the East China Mountains, the Qinghai66 156
54 266
(6.8%)
(5.4%)
Xizang Plateau, Hainan Island, and Taiwan
Endangered
Endemic
Endangered
Endemic
Island. According to the criteria of UNEPWCMC (2001), all of these ecoregions are
located in mountainous areas and should be
Figure 3. Venn diagrams illustrating the spatial congruence between different aspects considered as the hotspot ecoregions (with a
(overall, endemic, and endangered) of biodiversity hotspots for (a) plants and (b) animals. hotspot ecoregion determined by an area
For each of the aspects, the hotspots are the top 5% land areas with the highest density.
having at least 5% of its land area composed
of hotspot counties) for biodiversity in
Comparison between the different aspects of
China. These 10 biodiversity hotspot ecoregions cover an
biodiversity hotspots
area of ca 2.3 million km2, or 24.2 % of Chinas total land
The various types of biodiversity hotspot counties (those area. Table 2 illustrates the plant diversity of these areas.
based on the density of all, endemic, and endangered taxa) In total, they contained 3110 plant genera (92.0% of the
tended to be clustered in a few key regions such as the countrys total), 220 (90.5%) endemic plant genera, 366
Hengduan Mountains, the upper reaches of the Hongshui (94.3%) endangered plants, 427 (91.0%) terrestrial mamRiver, the East China Mountains, Hainan Island, and mal species, 65 (85.5%) endemic mammals, and 254
Taiwan Island (Figure 2). However, there remains some (72.2%) endangered vertebrates. It is worth noting that
spatial dissimilarity among the different types of hotspot the alpine areas of the Tibetan Plateau contain globally
counties: for example, although the co-occurrence of high unique ecosystems and their species are highly endemic,
richness of overall, endangered, and endemic plants and despite low overall taxonomic richness (Table 2).
animals was apparent in many of the same ecoregions,
These hotspot ecoregions are generally the same as
there was little plantanimal richness overlap, as they were those proposed by Chen (1998) and Olson and Dinerstein
often found in different counties within these ecoregions.
(1998), who used relatively descriptive and approximate
The three aspects of plant hotspot counties occupied a species information. Chen (1998), using estimated species
total area of 977 450 km2, of which only 11.5% (112 281 richness and number of endemics, proposed 11 terrestrial
km2) was common to all three measures (Figure 3a). An critical regions for the conservation of Chinas biodiveradditional 21.1% (206 025 km2) was shared between pairs sity, eight of which were located within the hotspot ecoreof aspects. However, 22% (215 497 km2), 23.7% (231 801 gions identified in this study. Of the WWFs Global 200
km2), and 21.7% (211 846 km2) were idiosyncratic to most critical and endangered ecoregions, 17 are located in
individual aspects of overall, endemic, and endangered or around China, with 11 consistent with the priority
hotspots, respectively (Figure 3a).
areas identified in this study. Several of the others are
Similarly, the animal hotspot counties also showed low either located in aquatic ecoregions or along Chinas borspatial congruence between different aspects. The cumula- ders (Olson and Dinerstein 1998).
tive area of the three types of animal hotspot counties was
1 001 709 km2, of which only 8.2% (82 080 km2) was com-  Conclusions
mon to all three aspects (Figure 3b). The idiosyncratic
hotspot counties occupied a total of 67.5% (675 977 km2), The patterns of biodiversity, as well as the mechanisms
with 20.0% (200 328 km2), 25.5% (254 852 km2), and 22% that cause these patterns, are a primary focus of biodiver(220 797 km2) for the overall, endemic, and endangered sity research (Gaston 2000). We explored the distribuanimal hotspots, respectively. The remaining 24.3% tion patterns of different aspects of biodiversity at a
(243 652 km2) was shared between pairs of the different county level, and identified ten biodiversity hotspot
aspects (Figure 3b). Similarly low spatial congruencies ecoregions that should serve as conservation priorities in
between different aspects of hotspots were observed with China. All of the identified hotspot ecoregions for overthe global pattern of avian richness (Orme et al. 2005).
all, endemic, and endangered diversity for plants and animals in China are located in the mountainous areas.
These mountainous areas are therefore critical for proTen hotspot ecoregions in major mountain ranges
tecting Chinas biodiversity.
High densities of plant genera and mammal species,
By the end of 2004, China had established more than
endemic plant genera and mammal species, and endan- 2000 national and local nature reserves, which collectively
(a) Plants

(b) Animals

Overall

The Ecological Society of America

Overall

www.frontiersinecology.org

351

Biodiversity in Chinas mountains

352

ZY Tang et al.

Table 2. Plant species richness of Chinas major mountain ranges and the Qinghai-Xizang (Tibetan) Plateau
Mountain range

Area
(103 km2)

Summit
Area (103 km2) of
hotpot counties (%) (m asl)

Seed plants
Genera/species
Reference*
(family/genus/species) endemic to China (local)

Hengduan Mts

~ 500

409 (81.7%)

7556

175/1348/7962

56/5079 (12/2988)

Wang 1994

Upper reaches of
Hongshui River
(Wumeng Mts)

110

76 (69.1%)

2900

178/1293/5061

55/2555 (8/468)

Wu 1996

East China Mts

~ 600

213 (35.5%)

2158

174/1180/4259

54/2147 (5/425)

Zhang and Lai 1993

Wuling Mts

~ 100

97 (96.8%)

2570

20/1005/4119

65/2682 (0/126)

Chen et al. 2002

Nanling Mts

45

44 (98.0%)

2142

200/1045/3854

41/na (2/na)

Chen and Zhang 1994

Taiwan Island

35

35 (100%)

3997

186/1201/3656

17/1275 (4/1070)

Ying and Hsu 2002

Qinling Mts

70

67 (96.1%)

3767

198/1007/3446

46/1428 (0/192)

Ying 1994

Xishuangbanna

20

20 (100%)

2500

197/1140/3336

8(0)/na (121)

Zhu et al. 2001

Hainan Island

35

35 (100%)

1867

204/1206/3315

19/1110 (8/505)

Xing et al. 1995

Alpine Qinghai Xizang Plateau

(> 4200 m)

~ 800

255 (31.9%)

8848

67/339/1816

14/636 (10/604)

Wu et al. 1995

*for detailed data source, see Web Only material; na = not available

cover approximately 1.7 million km2, or 17.8% of the countrys total land area (Chinese Species Information Service
2005). Of these, 1498 (or 73.8% of all of the protected
areas) are located in mountainous regions, covering a total
area of 1.5 million km2 (83.0% of all protected areas).
Although these reserves play a key role in Chinas plan for
biodiversity conservation, most are located in remote and
poorer areas and are often viewed as obstacles to human
welfare and economic development (Liu et al. 2003).
Therefore, maintaining the functionality of these reserves
is, and will continue to be, a critical but difficult issue.
Increasing financial support to reserves and local residents,
and assisting local communities in sharing the benefits
derived from tourism and natural resources are important
factors in addressing this problem (Liu et al. 2003).

 Acknowledgement
We are grateful to XP Wang, YH Chen, X Lin, W Xu, and
Y Chen for their assistance in data collection. This study
has been funded by the National Natural Science
Foundation of China (#49971002, #40501025,
#40024101, and #90211016) and the State Key Basic
Research and Development Plan (#G2000046801).

 References

Chen CD (Ed). 1998. Biodiversity of China: a country study.

Beijing, China: Environmental Science Press.
Chinese Species Information Service (CSIS). 2005. www.chinabiodiversity.com. Viewed 15 Feb 2005.
Dobson AP, Rodriguez JP, Roberts WM, and Wilcove DS. 1997.
Geographic distribution of endangered species in the United
States. Science 275: 55053.
www.frontiersinecology.org

Editorial Committee for the Administrative Division Atlas of

China. 2002. Administrative Division Atlas of China. Beijing,
China.
Editorial Committee for Chinas physical geography. 1985. Chinas
physical geography. Beijing, China: Science Press.
Fu LG. (Ed). 1992. China red data book of endangered plants (I).
Beijing China: Science Press.
Gaston KJ. 2000. Global patterns in biodiversity. Nature 405:
22027.
Krner C and Spehn EM. 2002. Mountain biodiversity: a global
assessment. New York, NY: The Parthenon Publishing Group.
Liu J, Ouyang Z, Pimm SL, et al. 2003. Protecting Chinas biodiversity. Science 300: 124041.
McNeely JA, Miller KR, Reid WV, et al. 1990. Conserving the
worlds biological diversity. Washington, DC: The World Bank.
Olson D and Dinerstein E. 1998. The Global 200: a representative
approach to conserving the Earths most biologically valuable
ecoregions. Conserv Biol 12: 50215.
Orme CDL, Davies RG, Burgess M, et al. 2005. Global hotspots of
species richness are not congruent with endemism or threat.
Nature 436: 101619.
Qian H. 1998. Large-scale biogeographic patterns of vascular plant
richness in North America: an analysis at the genera level.
J Biogeogr 25: 82936.
UNEPWCMC. 2002. Mountains and mountain forests. Cambridge, UK: World Conservation Monitoring Centre.
Wang S (Ed) 1998. China red data book of endangered animals.
Beijing, China: Science Press.
Wu ZY and Ding TY. 1999. Seed plants of China. Kunming,
China: Yunnan Science and Technology Press.
Zhang YZ (Ed). 1997. Distribution of mammalian species in China.
Beijing, China: China Forestry Publishing House.

Department of Ecology, College of Environmental Sciences, and Key

Laboratory for Earth Surface Processes of the Ministry of Education,
Peking University, Beijing, China, 100871 *(jyfang@urban.pku.edu.cn).
The Ecological Society of America