dynamic response of terrestrial hydrological cycles and...

Dynamic Response of Terrestrial Hydrological Cycles and Plant Water Stressto Climate Change in China

FULU TAO

Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, China

ZHAO ZHANG

State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing, China

(Manuscript received 6 May 2010, in final form 30 September 2010)

ABSTRACT

Rising atmospheric CO2 concentration CO2 and climate change are expected to have a major effect on

terrestrial ecosystem hydrological cycles and plant water stress in the coming decades. The present study

investigates the potential responses of terrestrial ecosystem hydrological cycles and plant water stress across

China to elevated CO2 and climate change in the twentieth and twenty-first centuries using the calibrated and

validated Lund–Potsdam–Jena dynamic global vegetation model (LPJ-DGVM) and eight climate change

scenarios. The spatiotemporal change patterns of estimated evapotranspiration (ET), soil moisture, runoff,

and plant water stress due to climate change and elevated CO2 are plotted singly and in combination. Positive

future trends in ET, soil moisture, and runoff—although differing greatly among regions—are projected.

Resultant plant water stress over China’s terrestrial ecosystem generally could be eased substantially through

the twenty-first century under the climate scenarios driven by emission scenarios that consider economic

concerns. By contrast, under the climate scenarios driven by emission scenarios that consider environmental

concerns, plant water stress could be eased until 2060, then begin to fluctuate until 2100. The net impact of

physiological and structural vegetation responses to elevated CO2 could result in an increasing trend in runoff

in southern and northeastern China, and a decreasing trend in runoff in northern and northwestern China in

the twentieth century. It is projected to reduce ET by 1.5 3 109 to 6.5 3 109 m3 yr21 on average, and increase

runoff by 1.0 3 109 to 5.4 3 109 m3 yr21 during 2001–2100 across China’s terrestrial ecosystems, although the

spatial change pattern could be quite diverse. These findings, in partial contradiction to previous results,

present an improved understanding of transient responses of China’s terrestrial ecosystem hydrological cycles

and plant water stress to climate change and elevated CO2 in the twentieth and twenty-first centuries.

1. Introduction

The climate and ecosystem interact with each other

through energy, water, and carbon exchange. Temper-

ature, radiation, and water interact to impose complex

and varying limitations on vegetation dynamics in dif-

ferent parts of the world (Churkina and Running 1998).

Among the major limiting factors to plant growth, water

availability most strongly limits vegetation growth over

40% of the earth’s vegetated surface (Nemani et al. 2003).

In Asia, the spatial and temporal variability of ecosystems

are characterized by their strong response to variations in

monsoon rainfall (Fu and Wen 1999). Particularly, var-

iability in precipitation regimes at seasonal and longer

time scales strongly influences ecosystem dynamics in

arid and semiarid regions (Lotsch et al. 2003; Tao et al.

2008).

Rising atmospheric CO2 concentration [CO2], land use

change, and climate change are expected to have a ma-

jor effect on global hydrological cycles, water resource

availability, and terrestrial ecosystem dynamics in the

coming decades (e.g., Alcamo et al. 2003; Tao et al. 2003;

Arnell 2004; Barnett et al. 2005; Milly et al. 2005; de Wit

and Stankiewicz 2006; Oki and Kanae 2006; Gedney et al.

2006; Gerten et al. 2007; Piao et al. 2007; Koutsoyiannis

et al. 2007; Alo and Wang 2008). Climate change directly

influences precipitation patterns, the hydrological cycle,

and vegetation dynamics, leading to substantial changes

in land surface dynamics (e.g., Milly et al. 2005; Alo and

Corresponding author address: Fulu Tao, Institute of Geographic

Sciences and Natural Resources Research, Chinese Academy of

Sciences, Beijing 100101, China.

E-mail: [email protected]

JUNE 2011 T A O A N D Z H A N G 371

DOI: 10.1175/2010JHM1314.1

� 2011 American Meteorological Society

Wang 2008; Tao et al. 2008). On the other hand, changes in

land surface properties—such as albedo, surface rough-

ness, rooting depth, stomatal resistance, and leaf area

index (LAI)—can feed back on the hydrological cycle

and climate through the energy and water fluxes (e.g.,

Sellers et al. 1997; Beltran-Przekurat et al. 2008). Under

elevated [CO2], stomatal apertures on plant leaves are

observed to close partially—suppressing transpiration

and, consequently, increasing runoff—together with pre-

cipitation increase (e.g., Gedney et al. 2006; Betts et al.

2007); structural changes in vegetation in response to

increased productivity, particularly changes in LAI, can

also alter canopy interception and transpiration (e.g.,

Sellers et al. 1997; Cowling and Field 2003; Piao et al.

2007; Alo and Wang 2008).

Numerous studies have investigated the trends in

river runoff and water availability for human use (e.g.,

Vorosmarty et al. 2000; Alcamo et al. 2003; Arnell

2004; Barnett et al. 2005; Milly et al. 2005; de Wit and

Stankiewicz 2006) and changes in terrestrial ecosystem

hydrological cycles due to climate change and/or human

activities on a global scale (e.g., Tao et al. 2003; Gedney

et al. 2006; Gerten et al. 2007; Piao et al. 2007; Alo and

Wang 2008). The results generally show that climate

change, land use change, and [CO2] have substantial ef-

fects on terrestrial ecosystem hydrological cycles, which

are expected to accelerate in a warmer future climate.

Increasing evapotranspiration (ET), deceasing runoff,

and increasing drought risk in some regions of the world,

accompanied with pronounced increases in precipita-

tion and runoff in other regions, have been indicated.

However, the mechanisms underlying the observational

runoff changes, particularly the roles of physiological

(e.g., stomatal) and structural (e.g., LAI) responses of

vegetation to elevated [CO2], remain uncertain. For

example, Gedney et al. (2006), using a mechanistic land

surface model, attributed increases in continental runoff

in the twentieth century to rising [CO2] via the suppres-

sion of transpiration. Piao et al. (2007), using a process-

based terrestrial biosphere model [Organizing Carbon

and Hydrology in Dynamic Ecosystems (ORCHIDEE)]

accounting for both the physiological and structural re-

sponse of elevated [CO2], found a decrease in global

runoff in the twentieth century. Gerten et al. (2007)

explored the effects of rising [CO2] and climate change

on soil moisture and biome-level water limitation during

2071–2100 using the Lund–Potsdam–Jena dynamic global

vegetation model (LPJ-DGVM) forced by five different

climate scenarios based on the same Intergovernmental

Panel on Climate Change (IPCC) 1992 emission scenario

(IS92a). They found that ecosystems in northern tem-

perate latitudes were at greatest risk of increasing water

limitation, in contrast with most other latitudes, where

ecosystem water limitation tended to be eased. Alo

and Wang (2008) used a Community Land Model, ver-

sion 3 (CLM3) land surface model driven by eight cli-

mate scenarios to explore the hydrological impact of the

potential vegetation response to climate change in 2100.

They showed that climate- and CO2-induced changes in

potential vegetation structure substantially influenced

the surface hydrological processes.

Dynamic responses of hydrological cycles and vege-

tation water stress at a regional scale to climate change

are complex, with both regional increases and decreases

expected in runoff, drought, and flooding risk (e.g., Tao

et al. 2005; de Wit and Stankiewicz 2006; Barnett et al.

2008; Vano et al. 2008; Mishra et al. 2010). In China,

numerous studies have investigated the responses of

streamflow to changes in climate and/or land use/cover

in some watersheds (e.g., Zhang and Wang 2007; Zhang

et al. 2008; Cong et al. 2009). Less attention has been

paid to changes in terrestrial hydrological cycles and

water availability for natural ecosystems. Tao et al. (2005)

developed a terrestrial ecosystem water-balance model to

assess the interactive impacts of climate change and soil

degradation on terrestrial water cycles in China. Recent

studies highlight the importance of vegetation structure

feedback on the water balance of the land surface (e.g.,

Gedney et al. 2006; Gerten et al. 2007; Piao et al. 2007;

Alo and Wang 2008), suggesting that our understand-

ing of the dynamic response of terrestrial hydrological

cycles to climate change and rising [CO2] across China’s

terrestrial ecosystem could be improved with dynamic

global vegetation models (DGVMs).

In a companion paper (Tao and Zhang 2010), the LPJ-

DGVM (Sitch et al. 2003), with the updated hydrology

of Gerten et al. (2004), was calibrated and validated, and

forced by eight climate scenarios to investigate the dy-

namic responses of the terrestrial ecosystem’s structure

and function to transient change in [CO2] and climate

in China. In the present study, using the calibrated and

validated LPJ-DGVM over China, we aim to investigate

the dynamic responses of major components of terres-

trial hydrological cycles and vegetation water stress across

China’s terrestrial ecosystem to transient change in [CO2]

and climate in the twentieth and twenty-first centuries.

2. Methods and data

a. LPJ-DGVM

The LPJ-DGVM (Sitch et al. 2003) simulates photo-

synthesis, plant distribution, and competition of 10 plant

functional types (PFTs), including tropical broad-leaved

evergreen (TrBE), tropical broad-leaved rain green

(TrBR), temperate needle-leaved evergreen (TeNE),

372 J O U R N A L O F H Y D R O M E T E O R O L O G Y VOLUME 12

temperate broad-leaved evergreen (TeBE), temperate

broad-leaved summer green (TeBS), boreal needle-leaved

evergreen (BoNE), boreal needle-leaved summer green

(BoNS), boreal broad-leaved summer green (BoBS),

temperate herbaceous (TeH), and tropical herbaceous

(TrH). The current version includes the updated hy-

drology of Gerten et al. (2004). Human land use such

as cropland is not considered—all simulations reported

here are for potential natural vegetation. The plant dis-

tribution is based on bioclimatic limits for plant growth

and regeneration. PFT-specific parameters govern com-

petition for light and water among the PFTs. Dispersal

processes are not explicitly modeled, and an individual

PFT can invade new regions if its bioclimatic limits and

competition with other PFTs allow establishment. Car-

bon is stored in seven PFT-associated pools representing

leaves, sapwood, heartwood, fine roots, a fast and a slow

decomposing above-ground litter pool, and a below-

ground litter pool; in addition, two soil carbon pools for

each grid cell receive input from the litter pools of all

PFTs present. Photosynthesis is a function of absorbed

photosynthetic-active radiation, temperature, atmospheric

[CO2], day length, canopy conductance, and biochemical

pathways (C3 and C4), using a form of the Farquhar

scheme (Farquhar et al. 1980; Collatz et al. 1992) with

leaf-level optimized nitrogen allocation (Haxeltine and

Prentice 1996) and an empirical convective boundary

layer parameterization (Monteith 1995) to couple car-

bon and water cycles. Decomposition rates of soil and

litter organic carbon depend on soil temperature (Lloyd

and Taylor 1994) and moisture (Foley et al. 1996). The

model has been extensively used to study past, present,

and future terrestrial ecosystem carbon and hydrologi-

cal dynamics, as well as biochemical and biophysical

interactions between ecosystems and the atmosphere at

global, regional, and site scales (e.g., Cramer et al. 2001;

Bachelet et al. 2003; Sitch et al. 2003; Morales et al. 2007;

Scholze et al. 2006; Gerten et al. 2007; Sitch et al. 2008;

Doherty et al. 2010).

The water balance in the LPJ-DGVM is coupled to

vegetation dynamics and photosynthesis. The model dis-

tinguishes 2 soil layers with fixed thickness (upper, 50 cm;

lower, 100 cm). The soil moisture of each layer is up-

dated at daily time steps according to the previous day’s

water content and the balance between the amount of

water infiltrating into the soil and that removed from the

soil layers through runoff, percolation, and actual ET.

The ET is the sum of interception loss and plant tran-

spiration for all PFTs, plus evaporation from bare soil.

Transpiration is modeled as the lesser of an atmosphere-

controlled demand function and a plant-controlled sup-

ply function, as proposed by Federer (1982). The model

diagnoses surface and subsurface runoff from the excess

of water over field capacity of the upper and the lower soil

layer, respectively. The amount of water percolating

through the second soil layer is assumed to contribute to

subsurface runoff. There is no lateral redistribution of

water among grid cells and no routing in the stream

network. Soil water limitation restricts the transpiration

and photosynthesis rates, which emulates the decrease

of photosynthetic activity as the soil water content de-

clines and finally reaches the wilting point. The ratio

between actual and potential canopy conductance, con-

sequently, describes the extent to which transpiration and

photosynthesis are co-limited by soil water deficit. The

water balance computed by LPJ-DGVM has been com-

prehensively validated, and it was found that the model

performs on the level of state-of-the-art global hydro-

logical models with respect to the quality of runoff and

ET simulations (Gerten et al. 2004).

b. Model input data and modeling protocol

1) MODEL INPUT DATA

Model input data are monthly mean air temperature,

total precipitation and percentage of full sunshine, an-

nual atmospheric [CO2], and soil texture class. Monthly

fields of mean temperature, precipitation, cloud cover,

and the number of wet days on a 0.58 3 0.58 grid from

1901 to 2000 were taken from the Climatic Research

Unit (CRU) dataset called CRU time series (TS) 2.1

(Mitchell and Jones 2005). To get daily values used in

LPJ-DGVM, monthly mean temperature and percent-

age of full sunshine data are interpolated to provide

‘‘quasi-daily’’ time series of climate using linear inter-

polation; monthly precipitation is interpolated to daily

values using a weather generator, with monthly total

precipitation and wet days as inputs, after Gerten et al.

(2004). The dataset of historical global atmospheric

[CO2] was from the Carbon Cycle Model Linkage Project

(Kicklighter et al. 1999; McGuire et al. 2001). The da-

taset was extended to 2000 according to atmospheric

observation (Keeling and Whorf 2005). Soil texture

data were based on the Food and Agricultural Organi-

zation (FAO) soil dataset (Zobler 1986; FAO 1991).

For future climate and [CO2] scenarios, eight future

climate scenarios consisting of the combinations of two

GCMs and four IPCC Special Report on Emissions

Scenarios (SRES) emission scenarios (A1FI, A2, B1,

and B2) were used in the study. The two GCMs are third

climate configuration of the Met Office (UKMO) Uni-

fied Model (HadCM3) and Canadian Centre for Climate

Modelling and Analysis (CCCma) Coupled General

Circulation Model, version 2 (CGCM2). CGCM2 pre-

dicts warmer and drier conditions over China, while

HadCM3 predicts wetter conditions. The HadCM3


model includes CO2 effects on stomatal closure in its

land surface scheme, which affect the evolution of the

climate (Betts et al. 2007). This has been shown to

contribute to greater warming over land than would be

the case without CO2 physiological effects (Boucher

et al. 2009). By contrast, the land surface scheme in

CGCM2 uses a single soil layer with spatially varying

field capacity and soil properties. The scenarios on

monthly fields of mean temperature, precipitation, and

cloud cover on a 0.58 3 0.58 grid from 2001 to 2100 were

taken from the CRU at the University of East Anglia

(Mitchell et al. 2004). The monthly wet days in the

twenty-first century are assumed to replicate the ob-

served values in the twentieth century because the

variable is not available in the future climate change

scenarios. GCM outputs were interpolated to 0.58 res-

olution using a Delaunay triangulation of a planar set

of points. Climate scenarios of the twenty-first century

replicate observed month-to-month, interannual, and

multidecadal climate variability of the detrended twen-

tieth century climate. (The full method for its con-

struction is described in Mitchell et al. 2004.) Thus,

climate scenarios of the twenty-first century replicate

the observed month-to-month, interannual, and multi-

decadal climate variability of the detrended twentieth

century climate data. One consequence of this approach

is that possible future changes in multidecadal or in-

terannual variability from the GCMs are not included in

these scenarios. The dataset of global atmospheric CO2

concentration scenarios under A1FI, A2, B1, and B2

emission scenarios was taken from the IPCC report

(Prentice et al. 2001). For each year, the average of the

reference estimate from the Integrated Science Assess-

ment Model (ISAM) and the reference estimate from the

Bern Carbon Cycle (Bern-CC) model was used. The cli-

mate and CO2 scenarios used in the study are as in Tao

and Zhang (2010). In comparison with the baseline

(1961–90) level, atmospheric CO2 concentration is pro-

jected to change by a factor of 1.64 (B1) to 2.90 (A1FI) by

2100.

2) MODELING PROTOCOL AND ANALYSIS

The LPJ-DGVM has been calibrated to reasonably

capture the general patterns of vegetation distribution

across China’s terrestrial ecosystems. New parameter

values were used for bioclimatic limits, and the model

was validated against satellite, in situ, and inventory data

for net primary production (NPP), LAI, and carbon stor-

age simulations (Tao and Zhang 2010).

In the present study, the calibrated model was applied

to simulate the dynamic responses of terrestrial hydro-

logical cycles and plant water stress to climate change

and rising [CO2] across China from 1901 to 2100 at the

gridcell resolution of 0.58 3 0.58. The simulation for each

grid cell began from ‘‘bare ground,’’ and ‘‘spun up’’

(under nontransient climate) for 1000 model years to

develop equilibrium carbon pools and vegetation cover,

following the standard simulation by Sitch et al. (2003).

Then the model was driven with transient historical

climate (1901–2000) and each of the eight future climate

scenarios (2001–2100) from 1901 to 2100 together with

transient historical [CO2] (1901–2000) and future [CO2]

under the corresponding emission scenario (2001–2100)

(hereafter, ETRC experiment).

To quantify the hydrological impacts of climate change

and the direct physiological and structural effects of

[CO2] enrichment separately, we also have contrasting

simulation experiments with [CO2] held at year 2000

state from 2001 to 2100 for all 8 scenarios, while using

the same climate inputs as ETRC experiment (hereafter,

ETR experiment).

We further investigated the temporal and spatial

patterns and mechanisms in changes of major compo-

nents in terrestrial hydrological cycles such as ET, soil

moisture (Ws), and runoff, as well as the changes in plant

water stress, under eight different climate scenarios in

both the ETRC and ETR experiments. Direct physiological

and structural effects of [CO2] enrichment are derived as

the difference between the ETRC and ETR experiments.

c. Runoff and ET dataset for model validation

The University of New Hampshire (UNH)–World

Meteorological Organization’s Global Runoff Data Cen-

ter (GRDC) Composite Runoff Fields v1.0 (Fekete et al.

2002) are used to validate the simulated runoff. The

composite runoff fields are based on observed river dis-

charge and simulated water balances. Selected gauging

stations from the World Meteorological Organization

GRDC data archive were firstly geographically co-

registered to a gridded simulated topological network

at 0.58 spatial resolution. Interstation regions between

gauging stations along the network were identified and

annual interstation runoff was calculated. The annual

interstation runoff was compared with outputs from

water-balance model calculations, which were performed

using long-term mean monthly climate forcing (temper-

ature and precipitation). The simulated runoff for each

cell was multiplied by the ratio of observed to simulated

runoff of the corresponding interstation region from the

GRDC dataset to create spatially distributed runoff

fields at 0.58 spatial resolution.

The Global Hydrographic Dataset (GGHYDRO),

release 2.3.1 (Cogley 2003) is also used to compare with

simulated runoff. The runoff field has a spatial resolu-

tion of 18 longitude by 18 latitude. The runoff field

was derived from maps in a considerable number of


sources, of which Korzun et al. (1977) was the most

important.

A new version of the Moderate Resolution Imaging

Spectroradiometer (MODIS) ET product, developed by

Mu et al. (2007), is used to compare with simulated ET.

The ET product was developed using an algorithm based

on the Penman–Monteith method, using reanalysis me-

teorological data and 0.058 remote sensing data from

MODIS.

3. Results

a. Changes in temperature and precipitation

In comparison with the period 1961–90 (Figs. 1a,b),

during 2071–2100, under the HadCM3 A1FI scenario,

annual mean temperature would increase from 3.28 to

7.38C across China (Fig. 1c), with the lowest increase in

the southeast and the highest increase in the northeast

and northwest. Annual total precipitation would in-

crease in most of China by up to 880 mm, with the most

increase in the south (Fig. 1d). Under the HadCM3 B1

scenario, annual mean temperature would increase from

1.68 to 4.08C across China (Fig. 1e), with the lowest in-

crease in the southeast and the highest increase in the

northeast and northwest. Annual total precipitation

would increase in most of China by up to 520 mm, with

the highest increase in the south and southeast (Fig. 1f).

Under the CGCM2 A1FI scenario, annual mean tem-

perature would increase from 2.98 to 9.18C across China

(Fig. 1g), with the lowest increase in the southwest and

the highest increase in the northwest. Annual total pre-

cipitation would increase from 2410 to 1600 mm, and

decrease most in the southwest (Fig. 1h). Under the

CGCM2 B1 scenario, annual mean temperature would

increase from 1.28C to 3.98C across China (Fig. 1i), with

the lowest increase in the southeast. Annual total pre-

cipitation would increase from 2270 to 738 mm, and

decrease most in the southwest (Fig. 1j).

b. Changes in evapotranspiration

The LPJ-DGVM captures the general pattern of the

MODIS-based ET across China reasonably well, al-

though simulated ET is relatively smaller in some por-

tions of southeastern China and the Tibetan Plateau

(Fig. 2). Moreover, model estimates of ET can reasonably

capture the interannual changes of the MODIS-based

ET. For example, ET changes in portions of north-

eastern and southwestern China in 2000 (Figs. 2a,b), 2001

(Figs. 2c,d), and 2002 (Figs. 2e,f) are captured generally.

Simulations show that ET over China’s terrestrial

ecosystems has had an increasing trend since the 1920s.

Both ETRC (Fig. 3a) and ETR (Fig. 3b) experiments

indicate that ET could increase through the twenty-first

century, which could be ascribed to both climate and

vegetation structural responses to elevated [CO2]. In

comparison with the ETR experiment (Fig. 3b), the ETRC

experiment projects more consistent ET time series un-

der the eight climate change scenarios (Fig. 3a). The net

impact of physiological and structural vegetation re-

sponses to elevated [CO2] could reduce ET by 1.5 3 109

to 6.5 3 109 m3 yr21 averaged across China’s terrestrial

ecosystem during 2001–2100, depending on climate sce-

narios (Fig. 3c).

Spatially, simulated annual ET had a decreasing trend

in southern and southeastern China in the twentieth

century, but had an increasing trend in portions of

northern and northeastern China (Fig. 4a). Simulated

ET had a distinct pattern during 1961–90, with annual

mean ET more than 600 mm yr21 in southern and

southeastern China. By contrast, it was substantially

lower in northwestern and western China, approaching

zero in desert areas (Fig. 4b). The simulated ET pattern

is largely consistent with the spatial patterns of vegeta-

tion and climate (see also Tao and Zhang 2010). During

2071–2100, ETRC experiments indicate that annual mean

ET could decrease by up to 150 mm yr21 in southern

and southeastern China because of a decrease in solar

radiation and vapor pressure deficit and elevated [CO2]

under the HadCM3 A1FI scenario (Fig. 4c). By contrast,

ET is projected to increase in the Tibetan Plateau be-

cause of an increase in temperature and precipitation.

Generally, the projected ET pattern under the HadCM3

B1 scenario is quite similar to that under the HadCM3

A1FI scenario, although with less an extent of change

(Fig. 4d). Nevertheless, in comparison with the HadCM3

A1FI scenario, ET is projected to decrease in potions of

northeastern China and increase in potions of south-

eastern China under the HadCM3 B1 scenario (Fig. 4d).

ET is projected to decrease at most by up to 200 mm yr21

in southwestern China under the CGCM2 A1FI sce-

nario (Fig. 4e), whereas precipitation is projected to

decrease by up to 300 mm yr21 (Fig. 1h). Under the

CGCM2 B1 scenario, ET change is projected to have

a similar spatial pattern although less extensive (Fig. 4f).

c. Changes in soil moisture

The LPJ-DGVM distinguishes two soil layers with

fixed thickness. Relative water content (Ws) of the top

layer—expressed as the fraction of soil water content in a

day compared to plant-available water capacity (Wmax)—

is focused on here. Simulated mean Ws over China’s

terrestrial ecosystem had an increasing trend in the

twentieth century (Fig. 5a). The ETRC experiments indi-

cate that Ws could increase substantially under HadCM3

A1FI and A2 climate scenarios, and increase moderately


FIG. 1. Spatial patterns of annual mean (a) temperature and (b) precipitation during 1961–90, as well as the

differences in (c),(e),(g),(i) temperature and (d),(f),(h),(j) precipitation between the periods of 2071–2100 and 1961–

90 under (c),(d) HadCM3 A1FI; (e),(f) HadCM3 B1; (g),(h) CGCM2 A1FI; and (i),(j) CGCM2 B1 climate scenarios,

respectively.


under HadCM3 B1 and B2 and CGCM2 A1FI and A2

climate scenarios, although with considerable interannual

variability. However, under CGCM2 B1 and B2 climate

scenarios, simulated mean Ws could almost stay the

same as the present level (Fig. 5a). The ETR experiments

indicate that Ws could have a decreasing trend under

CGCM2-based climate scenarios (Fig. 5b). The net im-

pacts of elevated [CO2] could increase mean Ws by up to

0.04 at the end of the twenty-first century (Fig. 5c).

Spatially, simulated Ws had an increasing trend in

northwestern and northeastern China in the twentieth

century, but had a decreasing trend in the Tibetan Pla-

teau and southwestern China (Fig. 6a). During 1961–90,

simulated annual means of Ws differed greatly among

regions, with Ws . 0.50 in southern and southeastern

China, and Ws , 0.30 in the arid and semiarid regions

in northern and northwestern China (Fig. 6b). During

2071–2100, under HadCM3 A1FI and B1 climate

scenarios, Ws is projected to increase by up to 0.3 in large

portions in northern, northeastern, and southern China

(Figs. 6c,d). By contrast, under CGCM2 A1FI and B1

climate scenarios, Ws is projected to decrease by up to

0.3 in portions of western and southwestern China, al-

though with the largest increase located in the Tibetan

Plateau (Figs. 6e,f). The spatial pattern of Ws is com-

prehensively controlled by the spatiotemporal distribution

and interaction of temperature, precipitation, radiation,

vegetation, soil texture, and [CO2].

d. Changes in runoff

Runoff occurs if Ws exceeds Wmax in a day in a soil

layer. On annual and longer time scales, continental

river runoff is approximately equal to the difference

between land precipitation and ET. Runoff here is the

sum of surface runoff (runoff from the upper soil layer),

subsurface runoff (drainage from the lower soil layer),

FIG. 1. (Continued)


FIG. 2. Comparison between model estimates of (a),(c),(e) annual ET and (b),(d),(f) MODIS-based annual ET by Mu et al. (2007) in

(a),(b) 2000; (c),(d) 2001; and (e),(f) 2002.


and the amount of water percolating through the lower

soil layer. The mean field of simulated runoff during the

period 1961–90 (Fig. 7a) agrees reasonably well with the

UNH–GRDC Composite Runoff Fields, version 1.0

(Fekete et al. 2002) (Fig. 7b) and the runoff field based

on GGHYDRO (Cogley 2003) (Fig. 7c). However, the

model underestimates runoff in portions of southwest-

ern China. Across China’s terrestrial ecosystem, the

model captures the seasonal dynamics of runoff fairly

well, although it underestimates each month’s runoff in

comparison with the estimates based on UNH–GRDC

Composite Runoff Fields, version 1.0 (Fig. 7d). Among

other reasons, the model underestimation of runoff

is ascribed to human interventions; for example, the

contributions of forest clearing and croplands expansion

to runoff, which are not included in the simulations. The

ETRC experiments indicate that annual mean runoff over

China’s terrestrial ecosystems had an increasing trend in

the twentieth century (Fig. 8a), which is consistent with

the previous reconstruction of global runoff data (Labat

et al. 2004), although runoff was claimed to decline over

southeastern Asia during 1948–2004 because of de-

creases in precipitation (Dai et al. 2009). It is projected

to increase notably through the twenty-first century by

both ETRC and ETR experiments, except those driven

by CGCM2 B1 and B2 climate scenarios (Figs. 8a,b).

The net impact of physiological and structural vegetation

responses to elevated [CO2] could lead to an increase

in annual mean runoff by 1.0 3 109 to 5.4 3 109 m3 yr21

over China’s terrestrial ecosystems during 2001–2100,

depending on climate scenarios (Fig. 8c).

Spatially, runoff had an increasing trend in portions of

southern and northeastern China in the twentieth cen-

tury, but had a decreasing trend in portions of south-

western China (Fig. 9a). The resultant spatial pattern of

the trend in runoff is generally consistent with Piao et al.

(2007). During 1961–90, the regions with annual mean

runoff more than 250 mm yr21 were mainly in southern

China. By contrast, in the arid and semiarid regions of

northern and northwestern China, annual mean runoff

approaches 0 (Fig. 7a). During 2071–2100, under the

HadCM3 A1FI (B1) scenario, annual mean runoff is

projected to increase mainly in southern and southeast-

ern China by over 400 mm yr21 (200 mm yr21) by ETRC

experiments (Figs. 9b,c). In contrast, under the CGCM2

A1FI (B1) scenario, annual mean runoff is projected to

decrease in large portions of southern and southwestern

China (Figs. 9d,e) by over 200 mm yr21 (200 mm yr21).

The runoff change pattern is quite consistent with pre-

cipitation change pattern, suggesting runoff changes are

primarily attributed to change in precipitation.

e. Changes in plant water stress

Soil water limitation simultaneously restricts both

transpiration and photosynthesis rates. In LPJ-DGVM,

the ratio between actual and potential canopy conduc-

tance (Gap) describes the extent to which transpiration

and photosynthesis are co-limited by soil water deficit.

The Gap ranges between 0 and 1, with high values in-

dicating low water stress. The Gap is calculated daily for

each PFT, and then aggregated into monthly and annual

means for the entire grid cell while weighting the PFT-

specific values according to their fractional coverage of

the grid cell. Simulated annual mean Gap over China’s

terrestrial ecosystem has had an increasing trend since

FIG. 3. Model estimates of potential changes in annual ET over

China’s terrestrial ecosystem in the twentieth and twenty-first

centuries in the (a) ETRC and (b) ETR experiments, as well as (c) the

direct physiological and structural effects of elevated [CO2] on ET.


1920, although with considerable interannual variability

(Fig. 10a). The ETRC experiments indicate that Gap could

increase substantially through the twenty-first century,

particularly using the climate scenarios driven by A1FI

and A2 emission scenarios (Fig. 10a). However, under

the climate scenarios driven by B1 and B2 emission

scenarios, annual mean Gap over China’s terrestrial

ecosystem is projected to increase until 2060 and

FIG. 4. Spatial pattern of trends in (a) estimated annual ET in the twentieth century and (b) estimated annual

mean ET during 1961–90, as well as its change during 2071–2100 in the ETRC experiment under (c) HadCM3 A1FI

and (d) HadCM3 B1, and (e) CGCM2 A1FI and (f) CGCM2 B1 climate scenarios. The initials in (b) represent the

representative grids for further analysis, which are referred to in Fig. 12.


then begin to fluctuate until 2100 (Fig. 10a). The ETR

experiments indicate that average Gap over China’s

terrestrial ecosystem could decrease notably, except

under HadCM3 A1FI and A2 climate scenarios (Fig.

10b). The net impact of physiological and structural

vegetation responses to elevated [CO2] upon Gap is

the largest under CGCM A1FI and A2 scenarios—

the warmer and drier climate scenarios (Fig. 10c)—

suggesting plant responses to elevated [CO2] be greater

in a drought environment. Elevated [CO2] could in-

crease mean Gap over China’s terrestrial ecosystem by

0.03–0.10 at the end of the twenty-first century depending

on climate scenarios.

Spatially, simulated annual mean Gap had an increas-

ing trend in northern and northeastern China in the

twentieth century, but had a decreasing trend in portions

of southwestern China (Fig. 11a). During 1961–90, simu-

lated annual mean Gap varied considerably across China’s

terrestrial ecosystem (Fig. 11b). During 2071–2100, un-

der HadCM3-driven climate scenarios, ETRC experi-

ments indicate that annual mean Gap could increase

substantially in the Tibetan Plateau and the north China

plain, suggesting vegetation could benefit from im-

provement in soil moisture status (Figs. 11c,d). By

contrast, in portions of northeastern China where

simulated annual mean Gap was larger than 0.60 during

1961–90, annual mean Gap could reduce by up to 0.5,

suggesting vegetation could become severely stressed

by water because of climate change. Under CGCM2-

driven climate scenarios, ETRC experiments indicate

the similar Gap change pattern, except that vegetation

in southwestern China could undergo more serious

water stress (Figs. 11e,f). In addition, the changes of Ws

are not always reflected in analogous changes in Gap

because of the difference in responsiveness between

them to the physiological CO2 effect (Gerten et al.

2007).

f. Response mechanisms of terrestrial hydrologicalcycles at subregions

To gain further insights into the possible response

mechanisms of hydrological cycles in different subre-

gions, we plot temperature and precipitation, and sim-

ulated time series of runoff, ET, Ws, and Gap in four

representative grids across China (Fig. 12). The four

grids include a grid centered on longitude 123.258 and

latitude 49.758 in northeastern China (NE grid), a grid

centered on longitude 90.758 and latitude 33.758 in the

Tibetan Plateau (TP grid), a grid centered on longitude

106.258 and latitude 29.258 in southwestern China (SW

grid), and a grid centered on longitude 115.258 and lati-

tude 25.258 in southeastern China (SE grid). The locations

of the grids are shown in Fig. 4b. Specific mechanisms act

on hydrological dynamics in each grid. For example, in

the ETRC experiment under the HadCM3 A1FI climate

scenario, in the NE grid, with semiarid and arid cli-

mates, ET and Ws increase gradually because of an in-

crease in temperature and precipitation, and runoff could

increase sharply after 2020 because of an increase in

precipitation. Soil moisture could increase, and plant

water stress could be eased gradually through the twenty-

first century (Fig. 12). In the TP grid, with an arid cli-

mate, Ws is projected to increase gradually; ET and

runoff are projected to increase sharply after around

2050 because of an increase in precipitation, when

FIG. 5. As in Fig. 3, but for mean relative soil moisture.


grasslands would colonize present barren areas because

of amelioration in temperature and water conditions

(Tao and Zhang 2010). In the SW and SE grids, with

a humid climate, temperature has increased sharply since

the 1980s and precipitation has increased gradually since

the 1960s. The model simulates a decrease in ET since

the 1940s due to a decrease in solar radiation, vapor

pressure deficit, and elevated [CO2], and runoff would

increase notably since the 1960s. The Ws would increase

and plant water stress would be eased gradually (Fig. 12).

FIG. 6. As in Fig. 4, but for mean relative soil moisture.


4. Discussion

a. Hydrological response of terrestrial ecosystems toclimate change in China

The study quantifies the potential response of terres-

trial ecosystem hydrological cycles and water stress in

China to climate change in the twentieth and twenty-

first centuries under eight climate change scenarios. The

contributions of elevated [CO2], climate change, and,

consequently, vegetation dynamics are explored in the

study. Our results indicate that plant water stress over

China’s terrestrial ecosystem could be substantially eased

through the twenty-first century under the climate sce-

narios driven by A1FI and A2 emission scenarios. By

contrast, under the climate scenarios driven by B1 and

B2 emission scenarios, it could be notably eased until

2060 and then begin to fluctuate until 2100. The spatial

pattern further indicates that vegetation water stress in

some arid and semiarid regions—such as the Tibetan

Plateau, the north China plain, and portions of the

northeast China plain—could be eased considerably

under A1FI emission scenarios. The regions with Gap

decreasing are mainly located in the regions with less

water stress during 1961–90, such as some portions

of southern and northeastern China. In arid and semi-

arid regions, transpiration is mainly determined by wa-

ter availability; a slight increase in precipitation could

lead to a notable increase in actual transpiration and,

consequently, Gap. However, in humid and semihumid

regions, changes in actual transpiration and Gap are

FIG. 7. Spatial pattern of (a) simulated annual mean runoff during 1961–90, (b) the runoff field from the UNH–

GRDC Composite Runoff Fields v1.0 (Fekete et al. 2002) at 0.58 spatial resolution (the blank points indicate missing

values), (c) the runoff field from the GGHYDRO (Cogley 2003) at 1.08 spatial resolution, and (d) comparisons

between simulated mean monthly runoff during 1961–90 and estimates from UNH–GRDC Composite Runoff Fields

v1.0.


simultaneously determined by changes in precipita-

tion, solar radiation, and temperature. We also find that

the projected changes in hydrological cycle compo-

nents are highly dependent on climate change sce-

narios. HadCM3 and CGCM2 project precipitation

changes with opposite signs in portions of southern

China, which are reflected in the changes in plant water

stress, vegetation fraction, and LAI by LPJ-DGVM

(Tao and Zhang 2010). These changes further affect

ET, soil moisture, and runoff. For example, in southern

China, the runoff is projected to increase during 2071–

2100 under the HadCM3 GCM-based climate scenario,

but decrease under the CGCM2 GCM-based climate

scenario.

The results of this present study stand in partial con-

tradiction to previous studies that suggest possible nega-

tive consequences for drought and flooding risk, as well

as other effects on the ecosystem caused by climate

change. The findings provide a new and promising per-

spective on the hydrological response of terrestrial eco-

systems to future climate change in China. With different

foci and using different climate scenarios and models,

previous studies showed somewhat different response

patterns and consequences that were dominated by neg-

ative impacts. For example, Tao et al. (2005) found that

under the combined impacts of climate change and soil

degradation, soil moisture deficits would increase most

in central, western, and southwestern China; surface

runoff would increase most in southeastern China. In

investigating the response of river runoff to climate

change by a hydrological model, Zhang and Wang (2007)

indicated that drought risk in northern China could fur-

ther increase, while flooding risk in southern China could

further increase; therefore the water-related risk would

be aggravated because of climate change (Lin et al. 2007;

Zhang and Wang 2007). Using an equilibrium atmospheric–

vegetation interaction model, Wu et al. (2010) showed that

in most ecoregions over China, moisture condition—

defined as the ratio of potential evapotranspiration to

precipitation—would become much more arid and in-

duce decreases in NPP. In comparison with previous

studies on a global scale, our results do not support the

view that ET could increase notably in southern China

because of change in climate and [CO2] (Alo and Wang

2008), or that runoff could decrease in southern China

under HadCM3-based scenarios (Alo and Wang 2008),

which may be partly due to the different emission sce-

nario used. Nevertheless, our results support the view

that runoff could decrease in southern China under

CGCM2-based scenarios (Alo and Wang 2008). Our

findings on change patterns of soil moisture and plant

water stress are generally consistent with Gerten et al.

(2007) under HadCM3-based scenarios; however, the

vegetation distribution (Tao and Zhang 2010), ET (Fig. 2),

runoff (Fig. 7), and, consequently, other components of

hydrological cycles are better described—particularly in

northeastern China and the Tibetan Plateau—in the

present study.

These studies differ with respect to climate change

and land use scenarios, models, and parameterization in

capturing ecosystem characteristics and in accounting

for the key response mechanisms to climate change. The

inconsistent results highlight DGVMs that account for

the dynamic response of vegetation distribution to cli-

mate change. These direct physiological and structural

effects of elevated [CO2] are important, particularly on

a time scale of decades to centuries.

FIG. 8. As in Fig. 3, but for mean runoff.


b. Physiological and structural effects of elevated[CO2] on terrestrial hydrological cycles

Vegetation physiological responses to elevated [CO2]

are well known; that is, stomata respond to elevated

[CO2] by partial closure. The structural vegetation re-

sponses to elevated [CO2] include PFT transition, changes

in fractional coverage, height, LAI, etc. Increases in

fractional coverage and LAI could increase canopy

interception, reducing soil moisture and runoff. The

FIG. 9. Spatial pattern of trends in (a) estimated annual mean runoff in the twentieth century and estimated annual

mean runoff change during 2071–2100, relative to 1961–90, in the ETRC experiment under (b) HadCM3 A1FI and (c)

HadCM3 B1, and (d) CGCM2 A1FI and (e) CGCM2 B1 climate scenarios.


resultant spatial pattern of the trend in runoff in the

twentieth century by the ETRC experiment is generally

consistent with Piao et al. (2007). However, our results

do not support the view of Piao et al. (2007) that runoff

had a decreasing trend due to [CO2] change around the

world. Piao et al. (2007) indicated that the increased

vegetation LAI in response to rising [CO2] provided a

greater cumulative surface area for canopy water tran-

spiration and interception, which could enhance ET and

offset the suppression of plant transpiration due to CO2-

induced stomatal closure, resulting in a decrease in

global runoff in the twentieth century. With [CO2] held

at the level of 1901 (i.e., 296.3 ppm) from 1901 to 2000,

our simulations indicate that runoff had a decreasing

trend in southwestern China and an increasing trend in

portions of southern and northeastern China due to

climate change only in the twentieth century (Fig. 13a),

which is also consistent with Piao et al. (2007). However,

the net impacts of physiological and structural vegeta-

tion responses to elevated [CO2] resulted in an in-

creasing trend in runoff in southern and northeastern

China, and a decreasing trend in runoff in northern and

northwestern China (Fig. 13b). Our findings suggest that

the relative impacts of physiological and structural veg-

etation responses to elevated [CO2] on runoff should

consider vegetation type, vegetation dynamics, fractional

cover, soil moisture condition, etc. In southern China

with humid climate, where TeBE is the dominant PFT

with higher fractional cover and LAI, physiological im-

pacts of rising [CO2] on stomatal conductance are rela-

tively larger. By contrast, in northern and northwestern

China with arid and semiarid climate, where TeBS, TeH,

and desert vegetation are the dominant PFTs with lower

fractional cover and LAI, vegetation structural responses

to elevated [CO2] are relatively larger.

Controlled experiments with plants growing in CO2-

rich environments have indeed indicated that LAI in-

creased sharply from low (200 ppm) to ambient (360 ppm)

[CO2] but tended to saturate at higher [CO2] (Cowling

and Field 2003). As a consequence, it is possible that the

continuous increase in [CO2] will probably lead to a de-

cline in transpiration in the future, as suggested by field

experiments (Nie et al. 1992). Our results show that the

net impact of physiological and structural vegetation re-

sponses to elevated [CO2] could reduce ET, increase

runoff, and increase Ws by up to 0.04 and Gap by 0.03–0.10

on average across China’s terrestrial ecosystem at the

end of the twenty-first century; this would significantly

ease plant water stress. The findings are generally in ac-

cordance with extensive experimental evidence (summa-

rized, e.g., by Amthor 1995) and with results from other

macroscale modeling studies (e.g., Cao and Woodward

1998; Levis et al. 2000; Cramer et al. 2001; Kergoat et al.

2002; Betts et al. 2004; Gerten et al. 2007). Nevertheless,

the exact magnitude of the positive net effect on water

limitation is debatable. The physiological response of

plants to elevated [CO2] as observed in short-term

controlled field or greenhouse experiments may differ

in real-world ecosystems as [CO2] increases into the

future because of possible nutrient (such as nitrogen)

limitations on photosynthesis and acclimatization of

plant physiological processes to elevated [CO2] and

temperature (Hungate et al. 2003; Norby and Luo 2004;

Rustad 2006). Furthermore, the scaling of leaf-level

FIG. 10. Model estimates of potential changes in annual mean

Gap over China’s terrestrial ecosystem in the twentieth and twenty-

first centuries in the (a) ETRC and (b) ETR experiments, as well as

(c) the direct physiological and structural effects of elevated [CO2]

on Gap.


experiments to the canopy and ecosystem scale remains

a challenge (Norby and Luo 2004). Thus, the represen-

tation of CO2 effects on photosynthesis and stomatal

conductance in DGVMs is a potential source of uncer-

tainty in the simulated vegetation structural changes

and, consequently, the simulated surface hydrological

impacts (Alo and Wang 2008).

c. Interactive effects of temperature, precipitation,and [CO2] on terrestrial hydrological cycles

Terrestrial ecosystem responses to multifactor global

change—including changes in temperature, precipita-

tion, and [CO2]—are regulated by complex and non-

linear processes. Using LPJ-DGVM and three other

FIG. 11. As in Fig. 4, but for mean Gap.


models, Luo et al. (2008) showed that modeled in-

teractive effects were mostly positive on transpiration

due to increases in temperature and precipitation, or

between increases in temperature and elevated [CO2].

This can be explained by the fact that warming increased

vapor pressure deficit and potential transpiration, while

increase in precipitation increased water availability.

When the two factors were combined, transpiration pro-

cesses were reinforced, leading to positive interactions

(Luo et al. 2008). Elevated [CO2] at the ambient tem-

perature increased canopy aerodynamic resistance and

reduced evapotranspiration (Bridgham et al. 1999). With

FIG. 12. (a),(b),(g),(h) Annual mean temperature and total precipitation; (c),(d),(i),(j) model estimates of runoff and ET; and

(e),(f),(k),(l) relative soil moisture and Gap in the ETRC experiment in four representative grids: (a),(c),(e) NE (49.758N, 123.258E);

(b),(d),(f) TP (33.758N, 90.758E); (g),(i),(k) SW (29.258N, 106.258E); and (h),(j),(l) SE (25.258N, 115.258E). The locations of the grids are

shown in Fig. 4b. The HadCM3 A1FI climate scenario is used during 2001–2100.


rising temperature, these effects of elevated [CO2] dras-

tically decreased (Homma et al. 1999), leading to a posi-

tive interaction between [CO2] and temperature. The

interaction explains why the ETRC experiment projects

more consistent ET time series under the eight climate

change scenarios (Fig. 3a) than the ETR experiment. The

other modeled positive interactive effects on transpi-

ration were between elevated [CO2] and a decrease in

precipitation, largely due to the fact that CO2-induced

reduction in transpiration was usually stronger in dry

than in wet environments (Housman et al. 2006). The

interaction explains that the net impact of physiolog-

ical and structural vegetation responses to elevated

[CO2] upon Gap is the largest under CGCM A1FI and

A2 scenarios—the warmer and drier climate scenarios

(Fig. 10c). The interactive effects on transpiration were

FIG. 12. (Continued)


negative between increase in temperature and decrease

in precipitation, or between elevated [CO2] and increase

in precipitation. With decrease in precipitation, water loss

via transpiration was mainly determined by water avail-

ability and may not be further stimulated by warming.

Thus, the effects of warming with decrease in precipitation

were less than the sum of the main effects of individual

factors (Luo et al. 2008). The interactions explain partly

the diverse change pattern of hydrological cycle compo-

nents. The variables—such as relative soil moisture, runoff,

and Gap—are affected comprehensively by temperature,

[CO2], and particularly precipitation, and are therefore

more dependent on GCMs and emission scenarios.

d. Uncertainties of the study

The LPJ-DGVM version used in the study considers

only potential natural vegetation and does not include

agriculture or managed forests or the effects of human

activities on the landscape. Human land use changes—

such as cropland expansion and forest clearing—can

substantially modify land surface dynamics and, conse-

quently, surface hydrological response. The results are

probably distorted in agricultural regions to some ex-

tent. This study utilized recent multiple climate scenar-

ios to represent the uncertainty in climate and [CO2]

impacts arising from two distinct uncertainty sources:

uncertainty in the future emissions of greenhouse gases

and uncertainty in climate modeling. However, the re-

sults may be limited by the particular DGVM model

used in this study. Model comparisons showed that, under

the more extreme projections of future environmental

change, the responses of the DGVMs diverge markedly.

In particular, large uncertainties are associated with the

responses of tropical vegetation to drought and boreal

ecosystems to elevated temperatures and changing soil

moisture status (Cramer et al. 2001; Sitch et al. 2008).

In addition, models that are fully coupled biosphere–

climate—accounting for the feedbacks from the bio-

sphere to the atmosphere such as effects of changing

vegetation distribution on regional climate and atmo-

spheric [CO2] (e.g., Levis et al. 2000; Claussen et al.

2001)—can improve the studies. For the climate sce-

narios of the twenty-first century, the raw projections

themselves (Johns et al. 1997; Gordon et al. 2000) were

not used here; rather further processed versions of the

data that are intended to deal with systematic biases and

provide downscaled information at higher resolution

were used. The variability from the past was imposed

on the top of long-term trends of future changes from

the GCMs rather than using the variability simulated

in the future simulations by the original GCMs. Although

the outputs of GCMs were interpolated using a Delaunay

triangulation of a planar set of points, the spatial patterns

of climate scenarios are still dominated by the resolution

of the GCMs.

5. Conclusions

The LPJ-DGVM, calibrated and validated for China’s

terrestrial ecosystem, is applied to investigate the tem-

poral and spatial changes in ET, runoff, soil moisture,

and the resultant limitation on photosynthesis and tran-

spiration of terrestrial vegetation over China’s terrestrial

ecosystem in the twentieth and twenty-first centuries un-

der eight climate scenarios. Our findings indicate that the

response of China’s terrestrial ecosystem hydrological

FIG. 13. Spatial pattern of trends in estimated annual mean

runoff in the twentieth century (a) due to climate change only and

(b) due to the net impact of physiological and structural vegetation

responses to elevated [CO2] only.


cycles to future climate change could be considerable.

Mean vegetation water stress on plant growth could be

generally eased by climate changes and rising [CO2],

although it could become more serious in portions of

northeastern China. The net impact of physiological and

structural vegetation responses to elevated [CO2] re-

sulted in an increasing trend in runoff in southern and

northeastern China and a decreasing trend in runoff in

northern and northwestern China in the twentieth cen-

tury. And it could generally lead to a decrease in mean

ET, as well as an increase in mean soil moisture and

runoff across China’s terrestrial ecosystem in the twenty-

first century, although the spatial change patterns could

be quite diverse. Our findings, in partial contradiction to

previous results, present an improved understanding of

transient responses in China’s terrestrial ecosystem hy-

drological cycles and plant water stress to climate change

and elevated [CO2] in the twentieth and twenty-first

centuries. The results are, however, constrained by the

lack of consideration for human land use changes and

vegetation feedback to climate, as well as the uncertainty

related to the physiological response of ecosystems to

elevated [CO2]. The future trajectories of ET, runoff, soil

moisture, and plant water stress are not to be taken as

predictions. Nevertheless, the overall temporal and spa-

tial patterns derived in the study can be used with some

confidence as a guide for policy making.

Acknowledgments. This study is supported by Na-

tional Key Program for Developing Basic Science (Pro-

ject 2009CB421105 and Project 2010CB950902), China.

F. Tao acknowledges the support of the ‘‘Hundred Tal-

ents’’ Program from the Chinese Academy of Sciences.

We are grateful to anonymous reviewers and editors for

their insightful comments on an earlier version of this

manuscript.

REFERENCES

Alcamo, J., P. Doll, T. Henrichs, F. Kaspar, B. Lehner, T. Rosch,

and S. Siebert, 2003: Global estimates of water withdrawals

and availability under current and future business-as-usual

conditions. Hydrol. Sci. J., 48, 339–348.

Alo, C. A., and G. Wang, 2008: Hydrological impact of the po-

tential future vegetation response to climate changes pro-

jected by 8 GCMs. J. Geophys. Res., 113, G03011, doi:10.1029/

2007JG000598.

Amthor, J. S., 1995: Terrestrial higher-plant response to increasing

atmospheric [CO2] in relation to the global carbon cycle.

Global Change Biol., 1, 243–274.

Arnell, N. W., 2004: Climate change and global water resources:

SRES emissions and socio-economic scenarios. Global Envi-

ron. Change, 14, 31–52.

Bachelet, D., and Coauthors, 2003: Simulating past and future

dynamics of natural ecosystems in the United States. Global

Biogeochem. Cycles, 17, 1045, doi:10.1029/2001GB001508.

Barnett, T. P., J. C. Adam, and D. P. Lettenmaier, 2005: Potential

impacts of a warming climate on water availability in snow-

dominated regions. Nature, 438, 303–309.

——, and Coauthors, 2008: Human-induced changes in the hy-

drology of the western United States. Science, 319, 1080–1083.

Beltran-Przekurat, A., C. H. Marshall, and R. A. Pielke Sr., 2008:

Ensemble reforecasts of recent warm-season weather: Impacts

of a dynamic vegetation parameterization. J. Geophys. Res.,

113, D24116, doi:10.1029/2007JD009480.

Betts, R. A., P. M. Cox, M. Collins, P. P. Harris, C. Huntingford,

and C. D. Jones, 2004: The role of ecosystem–atmosphere

interactions in simulated Amazonian precipitation decrease

and forest dieback under global climate warming. Theor.

Appl. Climatol., 78, 157–175, doi:10.1007/s00704-004-0050-y.

——, and Coauthors, 2007: Projected increase in continental runoff

due to plant responses to increasing carbon dioxide. Nature,

448, 1037–1041.

Boucher, O., A. Jones, and R. Betts, 2009: Climate response to the

physiological impact of carbon dioxide on plants in the Met

Office Unified Model HadCM3. Climate Dyn., 32, 237–249.

Bridgham, S. D., J. Pastor, K. Updegraff, T. J. Malterer, K. Johnson,

C. Harth, and J. Chen, 1999: Ecosystem control over

temperature and energy flux in northern peatlands. Ecol.

Appl., 9, 1345–1358.

Cao, M. K., and F. I. Woodward, 1998: Dynamic responses of

terrestrial ecosystem carbon cycling to global climate change.

Nature, 393, 249–252.

Churkina, G., and S. W. Running, 1998: Contrasting climatic con-

trols on the estimated productivity of global terrestrial biomes.

Ecosystems, 1, 206–215.

Claussen, M., V. Brovkin, and A. Ganopolski, 2001: Biogeophysical

versus biogeochemical feedbacks of large-scale land cover

change. Geophys. Res. Lett., 28, 1011–1014.

Cogley, J. G., 2003: GGHYDRO—Global Hydrographic Data,

release 2.3.1. Trent University Tech. Note 2003-1, 8 pp.

Collatz, G. J., M. Ribas-Carbo, and J. A. Berry, 1992: A coupled

photosynthesis–stomatal conductance model for leaves of C4

plants. Aust. J. Plant Physiol., 19, 519–538.

Cong, Z., D. Yang, B. Gao, H. Yang, and H. Hu, 2009: Hydro-

logical trend analysis in the Yellow River basin using a dis-

tributed hydrological model. Water Resour. Res., 45, W00A13,

doi:10.1029/2008WR006852.

Cowling, S. A., and C. B. Field, 2003: Environmental control of leaf

area production: Implications for vegetation and land-surface

modeling. Global Biogeochem. Cycles, 17, 1007, doi:10.1029/

2002GB001915.

Cramer, W., and Coauthors, 2001: Global response of terrestrial

ecosystem structure and function to CO2 and climate change:

Results from six dynamic global vegetation models. Global

Change Biol., 7, 357–373.

Dai, A., T. Qian, K. E. Trenberth, and J. D. Milliman, 2009:

Changes in continental freshwater discharge from 1948 to 2004.

J. Climate, 22, 2773–2792.

de Wit, M., and J. Stankiewicz, 2006: Changes in surface water

supply across Africa with predicted climate change. Science,

311, 1917–1921.

Doherty, R. M., S. Sitch, B. Smith, S. L. Lewis, and P. K. Thornton,

2010: Implications of future climate and atmospheric CO2

content for regional biogeochemistry, biogeography and eco-

system services across East Africa. Global Change Biol., 16,

617–640.

FAO, 1991: The Digitized Soil Map of the World, release 1.0. Food

and Agricultural Organization of the United Nations World


Soil Resources Rep. 67/1. [Available online at http://www.

fao.org/nr/land/soils/digital-soil-map-of-the-world/en/.]

Farquhar, G. D., S. von Caemmerer, and J. A. Berry, 1980: A

biochemical model of photosynthetic CO2 assimilation in leaves

of C3 species. Planta, 149, 78–90.

Federer, C. A., 1982: Transpirational supply and demand: Plant,

soil, and atmospheric effects evaluated by simulation. Water

Resour. Res., 18, 355–362.

Fekete, B. M., C. J. Vorosmarty, and W. Grabs, 2002: High-resolution

fields of global runoff combining observed river discharge and

simulated water balances. Global Biogeochem. Cycles, 16, 1042,

doi:10.1029/1999GB001254.

Flato, G. M., G. J. Boer, W. G. Lee, N. A. McFarlane, D. Ramsden,

M. C. Reader, and A. J. Weaver, 2000: The Canadian Centre

for Climate Modeling and Analysis global coupled model and

its climate. Climate Dyn., 16, 451–467.

Foley, J. A., I. C. Prentice, N. Ramankutty, S. Levis, D. Pollard,

S. Sitch, and A. Haxeltine, 1996: An integrated biosphere

model of land surface processes, terrestrial carbon balance, and

vegetation dynamics. Global Biogeochem. Cycles, 10, 603–628.

Fu, C. B., and G. Wen, 1999: Variation of ecosystems in East Asia

in association with the seasonal, interannual and interdecadal

variations of monsoon climate. Climatic Change, 43, 477–494.

Gedney, N., P. M. Cox, R. A. Betts, O. Boucher, C. Huntingford,

and P. A. Stott, 2006: Detection of a direct carbon dioxide

effect in continental river runoff records. Nature, 439,

835–838.

Gerten, D., S. Schaphoff, U. Haberlandt, W. Lucht, and S. Sitch,

2004: Terrestrial vegetation and water balance: Hydrological

evaluation of a dynamic global vegetation model. J. Hydrol.,

286, 249–270.

——, ——, and W. Lucht, 2007: Potential future changes in water

limitations of the terrestrial biosphere. Climatic Change, 80,

277–299.

Gordon, C., C. Cooper, C. A. Senior, H. Banks, J. M. Gregory,

T. C. Johns, J. F. B. Mitchell, and R. A. Wood, 2000: The

simulation of SST, sea ice extents and ocean heat transports in

a version of the Hadley Centre coupled model without flux

adjustments. Climate Dyn., 16, 147–168.

Haxeltine, A., and I. C. Prentice, 1996: BIOME3: An equilibrium

biosphere model based on ecophysiological constraints, re-

source availability, and competition among plant functional

types. Global Biogeochem. Cycles, 10, 693–709.

Homma, K., H. Nakagawa, T. Horie, H. Ohnishi, H. Y. Kim, and

M. Ohnishi, 1999: Energy budget and transpiration charac-

teristics of rice grown under elevated CO2 and high temper-

ature conditions as determined by remotely sensed canopy

temperatures. Japan J. Crop Sci., 68, 137–145.

Housman, D. C., E. Naumburg, T. E. Huxman, T. N. Charlet,

R. S. Nowak, and S. D. Smith, 2006: Increases in desert shrub

productivity under elevated carbon dioxide vary with water.

Ecosystems, 9, 374–385.

Hungate, B. A., J. S. Dukes, R. Shaw, Y. Luo, and C. B. Field, 2003:

Nitrogen and climate change. Science, 302, 1512–1513.

Johns, T. C., R. E. Carnell, J. F. Crossley, J. M. Gregory,

J. F. B. Mitchell, C. A. Senior, S. F. B. Tett, and R. A. Wood,

1997: The second Hadley Centre coupled ocean–atmosphere

GCM: Model description, spinup and validation. Climate Dyn.,

13, 103–134.

Keeling, C. D., and T. P. Whorf, cited 2005: Atmospheric CO2

records from sites in the SIO air sampling network. Trends: A

Compendium of Data on Global Change, Carbon Dioxide

Information Analysis Center, Oak Ridge National Laboratory,

U.S. Department of Energy. [Available online at http://cdiac.

ornl.gov/trends/co2/sio-keel.html.]

Kergoat, L., S. Lafont, H. Douville, B. Berthelot, G. Dedieu,

S. Planton, and J.-F. Royer, 2002: Impact of doubled CO2 on

global-scale leaf area index and evapotranspiration: Conflict-

ing stomatal conductance and LAI responses. J. Geophys.

Res., 107, 4808, doi:10.1029/2001JD001245.

Kicklighter, D. W., and Coauthors, 1999: A first-order analysis of

the potential role of CO2 fertilization to affect the global

carbon budget: A comparison study of four terrestrial bio-

sphere models. Tellus, 51B, 343–366.

Korzun, V. I., and Coauthors, 1977: Atlas of World Water Balance.

UNESCO Press, 34 pp. 1 65 maps.

Koutsoyiannis, D., A. Efstratiadis, and K. P. Georgakakos, 2007:

Uncertainty assessment of future hydroclimatic predictions:

A comparison of probabilistic and scenario-based approaches.

J. Hydrometeor., 8, 261–281.

Labat, D., Y. Godderis, J. L. Probst, and J. L. Guyot, 2004: Evi-

dence for global streamflow increase related to climate warm-

ing. Adv. Water Resour., 27, 631–642.

Levis, S., J. A. Foley, and D. Pollard, 2000: Large-scale vegetation

feedbacks on a doubled CO2 climate. J. Climate, 13, 1313–1325.

Lin, E., Y. Xu, S. Wu, H. Ju, and S. Ma, 2007: China’s National

Assessment Report on Climate Change (II): Climate change

impacts and adaptation (in Chinese). Adv. Climate Change

Res., 2, 51–56.

Lloyd, J., and J. A. Taylor, 1994: On the temperature dependence

of soil respiration. Funct. Ecol., 8, 315–323.

Lotsch, A., M. A. Friedl, B. T. Anderson, and C. J. Tucker, 2003:

Coupled vegetation–precipitation variability observed from

satellite and climate records. Geophys. Res. Lett., 30, 1774,

doi:10.1029/2003GL017506.

Luo, Y., and Coauthors, 2008: Modeled interactive effects of pre-

cipitation, temperature, and [CO2] on ecosystem carbon and

water dynamics in different climatic zones. Global Change

Biol., 14, 1986–1999.

McGuire, A. D., and Coauthors, 2001: Carbon balance of the ter-

restrial biosphere in the twentieth century: Analyses of CO2,

climate and land use effects with four process-based ecosystem

models. Global Biogeochem. Cycles, 15, 183–206.

Milly, P. C. D., K. A. Dunne, and A. V. Vecchia, 2005: Global

pattern of trends in streamflow and water availability in

a changing climate. Nature, 438, 347–350.

Mishra, V., K. A. Cherkauer, and S. Shukla, 2010: Assessment of

drought due to historic climate variability and projected future

climate change in the midwestern United States. J. Hydro-

meteor., 11, 46–68.

Mitchell, T. D., and P. D. Jones, 2005: An improved method of

constructing a database of monthly climate observations and

associated high-resolution grids. Int. J. Climatol., 25, 693–712.

——, T. R. Carter, P. D. Jones, M. Hulme, and M. New, 2004: A

comprehensive set of high-resolution grids of monthly climate

for Europe and the globe: The observed record (1901–2000)

and 16 scenarios (2001–2100). Tyndall Center for Climate

Change Research Working Paper 55, 30 pp.

Monteith, J. L., 1995: Accommodation between transpiring veg-

etation and the convective boundary layer. J. Hydrol., 166,

251–263.

Morales, P., T. Hickler, D. P. Rowell, B. Smith, and M. T. Sykes,

2007: Changes in European ecosystem productivity and car-

bon balance driven by regional climate model output. Global

Change Biol., 13, 108–122.


Mu, Q., F. A. Heinsch, M. Zhao, and S. W. Running, 2007: De-

velopment of a global evapotranspiration algorithm based on

MODIS and global meteorology data. Remote Sens. Environ.,

111, 519–536.

Nemani, R. R., C. D. Keeling, H. Hashimoto, W. M. Jolly,

S. C. Piper, C. J. Tucker, R. B. Myneni, and S. W. Running,

2003: Climate-driven increases in global terrestrial net pri-

mary production from 1982 to 1999. Science, 300, 1560–1563.

Nie, D., H. He, G. Mo, M. B. Kirkham, and E. T. Kanemasu, 1992:

Canopy photosynthesis and evapotranspiration of rangeland

plants under doubled carbon dioxide in closed-top chambers.

Agric. For. Meteor., 61, 205–221.

Norby, R. J., and Y. Luo, 2004: Evaluating ecosystem responses to

rising atmospheric CO2 and global warming in a multi-factor

world. New Phytol., 162, 281–294.

Oki, T., and S. Kanae, 2006: Global hydrological cycle and world

water resources. Science, 313, 1068–1072.

Piao, S., P. Friedlingstein, P. Ciais, N. de Noblet-Ducoudre,

D. Labat, and S. Zaehle, 2007: Changes in climate and land use

have a larger direct impact than rising CO2 on global river

runoff trends. Proc. Natl. Acad. Sci. USA, 104, 15 242–15 247.

Prentice, I. C., and Coauthors, 2001: The carbon cycle and atmo-

spheric carbon dioxide. Climate Change 2001: The Scientific

Basis, J. T. Houghton et al., Eds., Cambridge University Press,

185–224.

Rustad, L. E., 2006: From transient to steady-state response of

ecosystems to atmospheric CO2-enrichment and global cli-

mate change: Conceptual challenges and need for an in-

tegrated approach. Plant Ecol., 182, 43–62, doi:10.1007/

s11258-005-9030-2.

Scholze, M., W. Knorr, N. Arnell, and I. C. Prentice, 2006: A climate-

change risk analysis for world ecosystems. Proc. Natl. Acad. Sci.

USA, 103, 13 116–13 120.

Sellers, P. J., and Coauthors, 1997: Modeling the exchanges of

energy, water, and carbon between continents and the atmo-

sphere. Science, 275, 502–509.

Sitch, S., and Coauthors, 2003: Evaluation of ecosystem dynamics,

plant geography and terrestrial carbon cycling in the LPJ

dynamic global vegetation model. Global Change Biol., 9,

161–185.

——, and Coauthors, 2008: Evaluation of the terrestrial carbon

cycle, future plant geography and climate-carbon cycle feed-

backs using five dynamic global vegetation models (DGVMs).

Global Change Biol., 14, 2015–2039.

Tao, F., and Z. Zhang, 2010: Dynamic responses of terrestrial

ecosystem structure and function to climate change in China.

J. Geophys. Res., 115, G03003, doi:10.1029/2009JG001062.

——, M. Yokozawa, Y. Hayashi, and E. Lin, 2003: Terrestrial

water budget and the impacts of climate change. Ambio, 32,

295–301.

——, ——, ——, and ——, 2005: A perspective on water resources

in China: Interactions between climate change and soil deg-

radation. Climatic Change, 68, 169–197.

——, ——, Z. Zhang, Y. Hayashi, and Y. Ishigooka, 2008: Land

surface phenology dynamics and climate variations in the

North East China Transect (NECT), 1982–2000. Int. J. Remote

Sens., 29, 5461–5478.

Vano, J. A., J. A. Foley, C. J. Kucharik, and M. T. Coe, 2008:

Controls of climatic variability and land cover on land surface

hydrology of northern Wisconsin, USA. J. Geophys. Res., 113,

G04040, doi:10.1029/2007JG000681.

Vorosmarty, C. J., P. Green, J. Salisbury, and R. B. Lammers, 2000:

Global water resources: Vulnerability from climate change

and population growth. Science, 289, 284–288.

Wu, S., Y. Yin, D. Zhao, M. Huang, X. Shao, and E. Dai, 2010:

Impact of future climate change on terrestrial ecosystems in

China. Int. J. Climatol., 30, 866–873.

Zhang, J., and Q. Wang, Eds., 2007: The Impacts of Climate Change

on Hydrology and Water Resources. Science Press, 214 pp.

Zhang, X., L. Zhang, J. Zhao, P. Rustomji, and P. Hairsine, 2008:

Responses of streamflow to changes in climate and land use/

cover in the Loess Plateau, China. Water Resour. Res., 44,

W00A07, doi:10.1029/2007WR006711.

Zobler, L., 1986: A world soil file for global climate modeling.

NASA Tech. Memo. 87802, 32 pp.


dynamic response of terrestrial hydrological cycles and...

Documents