Content uploaded by Li Yu
Author content
All content in this area was uploaded by Li Yu on Nov 06, 2014
Content may be subject to copyright.
SPECIAL ISSUE
Li Yu ÆMingkui Cao ÆKerang Li
Climate-induced changes in the vegetation pattern of China
in the 21st century
Received: 4 February 2006 / Accepted: 10 August 2006 / Published online: 3 November 2006
ÓThe Ecological Society of Japan 2006
Abstract Quantifying climate-induced changes in vege-
tation patterns is essential to understanding land–cli-
mate interactions and ecosystem changes. In the present
study, we estimated various distributional changes of
vegetation under different climate-change scenarios in
the 21st century. Both hypothetical scenarios and Hed-
ley RCM scenarios show that the transitional vegetation
types, such as shrubland and grassland, have higher
sensitivity to climatic change compared to vegetation
under extreme climatic conditions, such as the evergreen
broadleaf forest or desert, barren lands. Mainly, the
sensitive areas in China lie in the Tibetan Plateau,
Yunnan-Guizhou Plateau, northeastern plain of China
and eco-zones between different vegetations. As the
temperature increases, mixed forests and deciduous
broadleaf forests will shift towards northern China.
Grassland, shrubland and wooded grassland will extend
to southeastern China. The RCM-project climate
changes generally have caused positive vegetation
changes; vegetation cover will probably improve 19%
relative to baseline, and the forest will expand to 8%
relative to baseline, while the desert and bare ground will
reduce by about 13%.
Keywords Climate change ÆVegetation pattern Æ
China Æ21st century
Introduction
Vegetation cover is an important factor in ecosystem
processes, such as water balance, energy exchanges,
nutrients and carbon cycles, and reacts strongly to cli-
mate change (Betts et al. 1997). The vegetation–climate
interaction in the changing global climate is one of the
most important issues in global change studies.
Compared to other counties, terrestrial ecosystems in
China will probably confront more risks to climate
change in the future because of its unique climatic
character and topography. Several models have assessed
the responses of vegetation and functions of ecosystems
based on bioclimatic classifications and processed-based
equilibrium terrestrial biosphere models (Alex and Pre-
ntice 1996; Bachelet et al. 2001; Neilson 1995; Smith
et al. 1992; Woodward 1995). Zhang used the Holdrige’s
life-zone scheme, which simulated the potential vegeta-
tion distribution and NPP pattern in China, as well as
analyzed the responses to climate change (Zhang et al.
1993). However, the Holdrige scheme as a statistical
approach has not been appreciated on the global scale
(Prentice et al. 1993; Chen et al. 2003; Yue et al. 2005).
Ni and Zhao used the BIOME3 and MAPSS models to
simulate the potential vegetation distribution and re-
sponses to climate change in China, respectively (Ni
et al. 2000; Zhao et al. 2002). They both simulated the
regime of the potential vegetation under baseline climate
and analyzed the responses of potential vegetation to
climate change; however, the two models were con-
nected with functions of ecosystems such as the NPP
or water useful efficient (WUE), so analyses for the
responses of potential vegetation were not the emphases
in their research. Gao and Yu based their work on a
regional dynamic vegetation model that simulated the
effects of climate change on potential vegetation distri-
bution (Gao et al. 2000). Vegetation distribution was
determined by spatiotemporal patterns of radiation
energy and available water resources in their model. As a
dynamic vegetation model, it considered the vegetation
Global changes in terrestrial ecosystems
L. Yu ÆM. Cao ÆK. Li
Institute of Geographic Sciences
and Natural Resources Research,
Chinese Academy of Sciences, Beijing,
People’s Republic of China
L. Yu (&)
National Climate Centre,
China Meteorological Administration,
Zhongguancun Nandajie 46#, Haidian,
Beijing 100081, People’s Republic of China
E-mail: yul@igsnrr.ac.cn
Tel.: +86-10-68408506
Fax: +86-10-68408758
Ecol Res (2006) 21: 912–919
DOI 10.1007/s11284-006-0042-8
migration rates, which were assumed to be proportional
to the gradient of green biomass. The dynamic vegeta-
tion model is an efficient tool that has reduced some of
the uncertainty about the responses of vegetation to
climate change. But in the early stages of development,
validating and capturing disturbance-related effects be-
come major challenges for the dynamic vegetation
models (Peng 2000). Different methods of climate–veg-
etation classifications and different sources of data, such
as plant function types, climatic data and soil data,
prevented the results from being comparable with each
other (Hurtt et al. 1998; Kittel et al. 1995). Table 1
compares several other studies. In general, based on the
existing research in China, the most explicit results are
that the transitional vegetation classes, like shrubs and
grasslands, are more sensitive to climate change than
other types of vegetation under extreme climatic condi-
tions. As for the general pole-ward shift of some forest
belts for specific vegetation types, each study had its own
views.
The present study was conducted to give a quantita-
tive analysis of the changing directions of the vegetation
pattern in China due to climate change. We used a
process-based ecosystem model to estimate the vegeta-
tion distribution and its response to climate change in
the 21st century.
Methods
Description of the model
To quantify climate-induced changes in vegetation dis-
tribution, we used the carbon exchange between vege-
tation, soil and the atmosphere (CEVSA) model that has
a vegetation module to describe vegetation distribution
and the corresponding primary productivity, carbon
stocks in biomass and litter production (Cao and
Woodward 1998a,b). In the model, mean temperatures
of the coldest and warmest month, growing degree days
above 0 and 5°C, and the annual soil moisture index (h)
are used to determine the vegetation types. The cold
tolerance of plant types determined by the minimum
mean temperature of the coldest month, the heat
requirement of plant types determined by annual accu-
mulated temperature over 0 or 5°C, and the wet index
indicated the water condition of the environment, which
determined the phenology of vegetation. A soil moisture
index (h), accounting for the extent of drought, is de-
fined as (Prentice et al. 1993):
h¼AET
D:
According to the concept of equilibrium evapo-
transpiration (Jarvis et al. 1986), evaporative demand D
is thus determined by the energy supply, i.e., the surface
net radiation.
The parameters determining distribution for different
vegetation types in our model are shown in Table 2as
follows. The other parameters in the CEVSA model are
detailed by Cao and Woodward (1998a,b). In China, for
the effects of monsoon and terrain, the character of
climate and vegetation differed from any other areas in
the world; it induced the unique distribution of vegeta-
tion. Consequently, the thresholds of climatic factors are
different from other regions (Ni et al. 2000; Zhao et al.
2002).
Data
The set of independent environmental variables needed
to run the CEVSA model consists of: temperature,
precipitation, cloudiness, relative humidity, soil texture
and atmospheric CO
2
concentration. The climatic and
soil data sets grid at a resolution of 0.1°latitude and 0.1°
Table 1 Comparison to the three main research studies of potential vegetation responses to climate change in China
Ni (2000) Zhao et al. (2002) Gao et al. (2000)
Model BIOME3 MAPSS Regional vegetation dynamic model
Scenarios Hadley GCM Had CM2 Hypothetic scenarios
a
Vegetation type classes 18 11 20
b
Functions NPP WUE NPP
Vegetation change
Expanded vegetation type Tropical deciduous forest
Grasslands
Shrubland
Steppe and savannas
Subtropical needleleaf/broadleaf mixed
forest
Tropical broadleaf evergreen forest
Shrubland
Evergreen needleleaf forest
Gramineal grass and short shrubs
Evergreen broadleaf forests
Temperate evergreen conifer forests
Deciduous broadleaf forest
Shrunk vegetation type Desert alpine tundra
Ice/polar desert
Boreal deciduous forest
Boreal forest
Tundra
Grassland
Desert
Savannas
Subtropical conifer forests
Deciduous shrubs
Deciduous conifer forests
Gramineal steppes
a
CO
2
concentration 100%, 2°C in monthly mean temperature, and 20% in monthly precipitation
b
Including three agricultural types
913
longitude. The data on the contemporary climate are
derived from the data of about 730 weather stations of
the Chinese Meteorological Administration. The atmo-
spheric CO
2
concentration used for the contemporary
climate was downloaded from the website of Manualoa
Observation. Soil data including soil types and soil
texture came from a digital map of soil texture in China.
The vegetation classification in our model used the
method from Hansen (2000). Under the classification
system, we simplified the natural vegetation into 12
types, which are: evergreen needleleaf forest, evergreen
broadleaf forest, deciduous needleleaf forest, deciduous
broadleaf forest, mixed forest, woodland, wooded
grassland, closed shrubland, open shrubland, grassland,
desert and barren. To distinguish the conversion of
vegetation due to climate change, we reclassified the 12
vegetation types into 4 general vegetations, which are
forest (including woodland), shrubland (including open
shrubland and closed shrubland), grassland and desert
(including barren land). In our prediction, we assumed
that the C:N ratio of the vegetation biomass does not
vary with the change in climate.
Climate change scenarios
To simulate the response of vegetation to climate change
in the 21st century, we used two approaches to drive the
CEVSA model, hypothetical scenarios and RCM sce-
narios. Based on most GCMs’ results in China, we de-
signed the probable climate change scenarios to simulate
the response of natural vegetation. We ran the model
with four combinatorial scenarios of temperature and
precipitation change that denoted the probable change
of the future climate. In detail, the four scenarios of
climatic change were ±20% precipitation, 2 and 4°C
temperature increase by the baseline, expressed with
-P20T2,-P20T4,P20T2 and P20T4, respectively. We
also used Hadley RCM A2 scenario data of the Hadley
Climate Centre to drive the CEVSA model. It gave
equilibrium climate scenarios with a resolution of 50 km
by 50 km, interpolated to a 0.1°latitude/longitude data
grid by Anusplin software. Relative to the baseline
(1961–1990) climate, the annual mean temperature rose
about 4.4°C, and annual precipitation changed between
330 and 780 mm.
Results
Simulation potential vegetation distribution
The model was integrated for 1961–1990 average cli-
matic factors as the baseline for simulating the potential
vegetation distribution under the contemporary climate.
The results showed that evergreen broadleaf forests were
mainly distributed in areas throughout southern China,
and then converted to mixed forest. In our vegetation
classification, mixed forest included deciduous/evergreen
mixed forest and broadleaf/needleleaf forest. Besides the
southeast of China, there was also a distribution of
mixed forest in northeast China. Moreover, deciduous
needleleaf forests were also distributed in areas
throughout northernmost China. In central China, from
east to west, vegetation covers were deciduous broadleaf
forest, woodland, wooded grassland, shrubland and
grassland, respectively, all pfs that are vegetation tran-
sition zones in China. In southwest China, the vegeta-
tion cover mainly consists of evergreen needleleaf forest.
However, in the south edge of the Tibetan Plateau, the
vegetation cover was complex. The vegetation cover
included evergreen broadleaf forests and grasslands. In
the northwest of China, most areas are desert and barren
with patches of grasslands or shrublands. Mixed forest
has the highest percentage of all vegetation types, about
23%, while other forest types totaled 23%. Desert and
barren land totaled 14 and 17%, respectively. The per-
cent of grassland was 11%. Savanna and shrubland were
recorded at around 6 and 3%, respectively.
The preliminary validation of our results compared
well with the result of BIOME3 by Ni et al. (2000) and
the map of Chinese vegetation by Hou (1982). The
regime of natural vegetation simulated by our model was
mostly consistent with their results; all the vegetation
types were roughly captured by the model. But it should
be explained that these comparisons were not qualitative
or site-specific comparisons. In this paper, our objectives
were to show the relationship between the main climatic
factors and potential vegetation in China. Therefore, it
could be used to simulate the potential response of
vegetation induced by climate change (Fig. 1).
Vegetation changes under the hypothetic scenarios
of climate change by the end of the 21st century
Figure 2and Table 3show the potential vegetation
distributions and the type percents of vegetation under
the four scenarios. The general regimes of vegetation
have comparability to the baseline, but the areas of each
vegetation type are very different under different climate
conditions. In general, evergreen broadleaf forests will
expend in the south of China, but deciduous needleleaf
Table 2 Climatic parameters for potential vegetation types in the
CEVSA model
Vegetation definition T
min
(°C) GD5 (°C) Wet
index
Evergreen/deciduous 18 3,200–4,500/1,100 0.65
Needleaf/broadleaf 2.5 – 0.55
Forest 25.0 1,100 0.50
Shrub – 850 0.30
Grass – 650 0.08
Desert – 550 –
Barran – – –
Note T
min
indicates minimum coldest month temperature; GD5
means accumulated temperature over 5°C
914
Fig. 1 The distribution of the
potential natural vegetation in
China simulated by the CEVSA
model
Fig. 2 Simulated distribution of vegetation in China under different hypothetic climatic scenarios
Table 3 The changed percent of vegetation types relative to baseline under the hypothetic scenarios
Vegetation types -P20T2 -P20T4 P20T2 P20T4
Evergreen needleleaf forest 10.3 1.3 30.3 50.9
Evergreen broadleaf forest 104.5 202.0 105.8 209.5
Deciduous needleleaf forest 55.4 67.3 38.5 48.2
Deciduous broadleaf forest 68.1 70.3 4.7 7.8
Mixed forest 42.0 71.3 15.2 45.2
Woodland 31.6 53.8 8.2 19.8
Wooded grassland 13.0 20.5 28.5 13.9
Closed shrubland 173.9 326.1 33.9 128.4
Open shrubland 155.4 176.3 116.0 124.0
Grassland 39.4 13.1 0.8 17.2
Desert 7.4 2.1 18.5 13.7
Barren 20.1 44.8 20.1 44.8
915
forest will shrink in the northeast of China due to tem-
perature increases. Shrubland will increase greatly and
barren land will decrease under the four climate change
scenarios.
The main shifts of vegetation due to precipitation
change take place between shrub and forest, grass and
shrub regardless of whether precipitation increases or
decreases. When precipitation increased, the shifts were
from shrubland to forest and from grassland to shrub-
land, which correlated with temperature increases. In
contrast, if precipitation decreased, the shifts were
mainly from forest to shrubland and from shrubland to
grassland.
Although the response of potential vegetation to
precipitation change and the shifts between vegetation
types were different than the response to temperature
change, the special regimes of sensitive areas to precip-
itation change resembled the temperature changes. Most
sensitive areas appeared in eco-transitions from forest to
grassland and occurred within the south brim of the
Tibetan Plateau. Correspondingly, precipitation de-
crease changed the regime of potential vegetation more
greatly than the effects of precipitation increase in Chi-
na.
The spatial changes of the four scenarios indicated
vegetation patterns that were high relative to precipita-
tion change (Figs. 3,4). Total areas affected were 40%
and almost 60% under two precipitation decrease sce-
narios. Moreover, the changed areas were 25 and 50%
under two precipitation increase scenarios, respectively.
In detail, under -P20T2 and -P20T4 scenarios, the most
change of vegetation occurred in forest shift to shrub-
land, shrubland to grassland and grassland to shrub-
land. In the north and northeast of China, most forest
will shift to shrubland. Original shrubland will shift
mostly to grassland. Vegetation cover on the Tibetan
Plateau will change dramatically. As a result, about 20%
of the precipitation will decrease and the temperature
will increase beyond 2°C of baseline, and forest areas
will decrease at least 16% under the baseline climate in
China. On the contrary, precipitation increase made a
more positive shift on vegetation, far more than the
negative shift. Forest areas will increase 43 and 52%
under the scenarios of P20T2 and P20T4 relative to the
forest areas of baseline. Besides some original shrubland
shift to forest, forest will be converted in the Tibetan
Plateau due to climate change. Profound changes in
grassland will also occur; most grassland will shift to
shrubland. From the four hypothetical scenarios, the
plains of north and northeast China were highly sensi-
tivity to precipitation change; the Tibetan Plateau and
eco-transition zones like shrubland were sensitive to
precipitation change and temperature change.
However, precipitation increase or decrease induced a
distinctness of vegetation regime, but the greatly chan-
ged areas were almost invariably to different climate
change scenarios. They mainly occured in the northeast
of China, the Tibetan Plateau, Yunnan-Guizhou Pla-
teau, transition areas of shrubland to grassland and
forest to shrubland. It indicated that the transitional
vegetation types, such as shrub and grassland, were
more sensitive to climate change than those under ex-
treme climate conditions, such as the evergreen broad-
leaf forest in the south of China and the deserts in the
northwest of China.
Fig. 3 The percent of vegetation under different hypothetic climatic
scenarios
Fig. 4 The shifts of vegetation relative to baseline under the
hypothetic scenarios
916
Vegetation changes under the Hadley RCM A2
scenarios by the end of the 21st century
Under the Hadley RCM A2 scenarios, evergreen
broadleaf forest became the dominant vegetation type
throughout most of southern China (Figs. 5,6). Ever-
green needleleaf forest increased in the southwest of
China, and most of the mixed forests and deciduous
broadleaf forests will spread throughout the northern
plain of China under the scenarios by the end of 21st
century. Relative to contemporary climate conditions,
evergreen forest will increase, especially evergreen
broadleaf forests. Deciduous forests and mixed forests
will also reduce, and deciduous needleleaf forests will
almost whither away completely in China. Shurbland
will increase, but grassland and desert will shrink to
some degree in the west of China, and barren land will
be reduced greatly. Vegetation cover will increase in the
Tibetan Plateau.
Under the climate of Hadley RCM A2 scenarios,
most vegetation types will change dramatically, but most
the change will be beneficial for vegetation cover. Most
shifts from grassland to shrubland will take place by the
end of 21st century; shrubland will shift to forest. The
negative shifts, i.e., shrublands to grasslands and forest
shift to shrubland, consisted of a small proportion of the
total vegetation change. The climate under Hadley
RCM A2 scenarios will enhance the cover of vegetation
in China. Forest areas will increase 58%, and the areas
of desert and bare ground will be reduced by about 32%
of their present areas (Table 4).
The main spatial change of vegetation occurred in the
Tibetan Plateau, Yun-Gui Plateau and the north and
northeastern plain of China under the Hadley RCM A2
Fig. 5 The spatial change of the vegetation regime relative to baseline under the hypothetical scenarios
Fig. 6 The distribution of
potential vegetation under
Hadley RCM A2 scenarios by
the end of the 21st century
917
scenarios, especially in the Tibetan Plateau. The results
indicated that these areas will have profound changes
according to the climate change and that current eco-
systems will be highly sensitive to climate change in the
future (Figs. 7,8,9). In some western areas of China,
climate change will improve the environmental condi-
tion, especially moist conditions, and vegetation cover
will increase in those areas. As a whole, climate condi-
tions will probably yield positive effects on the natural
vegetation in China without other disturbances. Total
vegetation cover increased about 19% under Hadley
RCM A2 scenarios relative to baseline.
Conclusions
We used the CEVSA model to quantify climate-induced
changes in vegetation distribution. Both the hypothetical
scenarios and the Hedley RCM scenarios showed that
the transitional vegetation ecosystems, such as shrub-
land and grassland, were more sensitive to climate
changes than other vegetation types under extreme cli-
matic conditions, such as the evergreen broadleaf forest
in the south of China and the desert and barren land in
the northwest of China. Main sensitive areas in China
are the Tibetan Plateau, Yunnan-Guizhou Plateau, the
northeastern plain of China, and those eco-zones be-
tween different ecosystems. Based on Hadley RCM
scenarios, by the end of the 21st century, vegetation
cover will probably improve in China, especially in the
Tibetan Plateau and some western areas of China, not
including other unforeseen disturbances. Forest areas
will increase by 8%, and desert and bare ground will
reduce by 13% of baseline. The total vegetation cover
will increase by 19% relative to the baseline climate. For
temperature increases, mixed forest and deciduous
broadleaf forest will shift towards northern China.
Grassland, shrubland and wooded grassland will extend
in the southeast of China.
Discussion
Our object for this paper is to analyze the effects induced
by climate change on the vegetation pattern. In fact,
different vegetation types have different time lags in
Table 4 The changed percent of vegetation types relative to base-
line under Hadley RCM A2 scenarios
Vegetation type Baseline Hadley RCM A2 Changed
Evergreen needleleaf forest 5.99 9.422 57.30
Evergreen broadleaf forest 8.288 25.371 206.12
Deciduous needleleaf forest 2.969 0.94 68.34
Deciduous broadleaf forest 4.612 5.569 20.75
Mixed forest 23.093 12.421 46.21
Woodland 1.282 1.511 17.86
Wooded grassland 6.463 5.815 10.03
Closed shrubland 2.839 10.603 273.48
Open shrubland 0.624 0.827 32.53
Grassland 11.971 8.756 26.86
Desert 14.084 9.184 34.79
Barren 17.783 9.583 46.11
0102030
1
2
3
4
5
6
7
10
9
8
11
12
Vegetation types
%
HadleyRCM
Baseline
Fig. 7 The percent of vegetation relative to baseline under Hadley
RCM A2 scenarios
Fig. 8 The shifts of vegetation relative to baseline under Hadley
RCM A2 scenarios
918
response to climate change and face varied kinds of
disturbances. The relationship between vegetation and
climate was not in equilibrium most of the time, so this
approach must include some uncertainties. However,
our analysis should still connect with the mechanics of
ecosystems and vegetation dynamics. Our work is a
primarily tentative approach to simulate the responses
of vegetation to climate change based on the relationship
between climate and vegetation, as well as to gain a
general comprehension of responses of potential vege-
tation to climate change in China.
Acknowledgments This research was supported by the Chinese
National Science Foundation.
References
Alex H, Prentice IC (1996) BIOME3: an equilibrium terrestrial
biosphere model based on eco-physiological constraints, and
competition among plant functional types. Glob Biogeochem
Cycles 4(10):693–708
Bachelet B, Neilson PR, James ML, Raymond JD (2001) Climate
change effects on vegetation distribution and carbon budget in
the United States. Ecosystems 4:164–185
Betts RA, Cox PM, Lee SE, Woodward FI (1997) Contrasting
physiological and structural vegetation feedbacks in climate
change simulations. Nature 387:796–799
Cao M, Woodward FI (1998a) Dynamic responses of terrestrial
ecosystem carbon cycling to global climate change. Nature
393:249–252
Cao M, Woodward FI (1998b) Net primary and ecosystem pro-
duction and carbon stocks of terrestrial ecosystems and their
responses to climate change. Glob Change Biol 4:185–198
Chen X, Zhang X, Li B (2003) The possible response of life zones in
China under global climate change. Glob Planet Change
38:327–337
Gao Q, Yu M, Yang X (2000) An analysis of sensitivity of ter-
restrial ecosystems in China to climatic change using spatial
simulation. Clim Change 47:373–400
Hansen MC, Defries RS, Townsgend JRG, Sohlberg R (2000)
Global land cover classification at 1 km spatial resolution a
classification tree approach. Int J Remote Sens 21(6/7):1331–
1364
Hou X (1982) Vegetation map of China (1:4000000), SinoMaps
Press, Beijing
Hurtt CG, Paul RM, Stephen WP, Simon AL (1998) Terrestrial
models and global change: challenges for the future. Glob
Change Biol 4:581–590
Jarvis PG, McNaughton KG (1986) Stomatal control of transpi-
ration: scaling up from leaf to region. Adv Ecol Res 15:1–49
Kittel TGF, Rosenbloom NA, Painter TH, Schimel DS et al (1995)
The VEMAP integrated database for modeling United States
ecosystem/vegetation sensitivity to climate change. J Biogeogr
22:857–862
Neilson PR (1995) Model for predicting continental-scale vegeta-
tion distribution and water balance. Ecol Appl 5(2):362–385
Ni J, Sykes TM, Prentice IC, Cramer W (2000) Modeling the
vegetation of China using the process-based equilibrium ter-
restrial biosphere model BIOME3. Glob Ecol Biogeogr 9:463–
479
Peng C (2000) From static biogeographical model to dynamic
global vegetation model: a global perspective on modeling
vegetation dynamics. Ecol Modell 135:33–54
Prentice IC, Cramer W, Harrison SP et al (1993) A global biome
model based on plant physiology and dominance, soil proper-
ties and climate. J Forest Biogeogr 19:117–134
Smith TM, Shugart HH, Bonan GB, Smith JB, Woodward FI
(1992) Modeling the potential response of vegetation to global
climate change. Adv Ecol Res 22:93–113
Woodward FI (1995) A global land primary productivity and
phytogeography model. Glob Biogeochem Cycles 9(4):471–490
Yue T, Fan Z, Liu J (2005) Changes of major terrestrial ecosystems
in China since 1960. Glob Planet Change. DOI 10.1016/j.glo-
placha.2005.03.001
Zhang X, Yang D (1993) A study on climate–vegetation interaction
in China: the ecological model for global change. Coenoses
8:105–119
Zhao M, Neilson PR, Yan X, Dong W (2002) Modeling the veg-
etation of China under changing climate. J Geogr Sci 57(1):28–
38
Fig. 9 The spatial change of
vegetation pattern under
Hadley RCM A2 scenarios
919
