10 a model for china’s energy requireme
description
Transcript of 10 a model for china’s energy requireme
Environmental Modelling & Software 22 (2007) 378e393www.elsevier.com/locate/envsoft
A model for China’s energy requirements and CO2 emissions analysis*
Ying Fan a, Qiao-Mei Liang a,b, Yi-Ming Wei a,*, Norio Okada c
a Institute of Policy and Management, Chinese Academy of Sciences, Beijing 100080, Chinab Graduate School, Chinese Academy of Sciences, Beijing 100080, China
c Kyoto University, Kyoto 611-0011, Japan
Received 17 December 2004; received in revised form 28 November 2005; accepted 9 December 2005
Available online 20 March 2006
Abstract
This paper introduces a model and corresponding software for modeling China’s Energy Requirements and the CO2 Emissions Analysis Sys-tem (CErCmA). Based on the inputeoutput approach, CErCmA was designed for scenario analysis of energy requirements and CO2 emissions tosupport policymakers, planners and others strategically plan for energy demands and environmental protection in China. In the system, majordrivers of energy consumption are identified as technology, population, economy and urbanization; scenarios are based on the major drivingforces that represent various growth paths. The inputeoutput approach is employed to compute energy requirements and CO2 emissions undereach scenario. The development of CErCmA is described in a case study: China’s energy requirements and CO2 emissions in 2010 and 2020 arecomputed based on the inputeoutput table of 1997. The results show that China’s energy needs and related CO2 emissions will grow exponen-tially even with many energy efficiency improvements, and that it will be hard for China to maintain its advantage of low per capita emissions inthe next 20 years. China’s manufacturing and transportation sectors should be the two major sectors to implement energy efficiency improve-ments. Options for improving this model are also presented in this paper.� 2006 Elsevier Ltd. All rights reserved.
Keywords: Energy requirement; CO2 emissions; Scenario analysis; Inputeoutput model
1. Introduction
Global warming, caused by increasing emissions of CO2
and other greenhouse gases as a result of human activities, isone of the major threats now confronting the environment.CO2 accounts for the largest share of total greenhouse gases,and its impact on the environment is also the greatest. If an-thropogenic CO2 emissions are allowed to increase withoutlimits, the greenhouse effect will further destroy the
* National Natural Science Foundation of China under grant Nos.70425001,
70573104 and 70371064, and the Key Projects of National Science and Tech-
nology of China (2001-BA608B-15, 2001-BA605-01).
* Corresponding author. Institute of Policy and Management (IPM), Chinese
Academy of Sciences (CAS), P.O. Box 8712, Beijing 100080, China. Tel.:
þ86 10 62650861; fax: þ86 10 62542619.
E-mail addresses: [email protected], [email protected] (Y.-M.
Wei).
1364-8152/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.envsoft.2005.12.007
environment for humans and all other living beings, threaten-ing the existence of humankind.
In order to control the continuous global warming and pro-tect the living environment, the Kyoto Protocol to the UnitedNations Framework Convention on Climate Changes, signedin Kyoto, Japan in 1997, sets detailed emissions mitigationcommitments for the 38 major industrialized countries.
Although the protocol did not set an explicit CO2 reductionobligation for China and other developing countries, thesenations still face great pressure from the environment. In2003, CO2 emissions caused by fuel combustion in Chinawere about 0.849 billion tons of carbon (tC), accounting for13.1% of the world’s total, second only to the United States,the largest CO2 emitter worldwide (IEA, 2003).
In addition to the current high CO2 emissions is the proba-bility that China’s economy will continue to grow rapidly overthe next 50 to 100 years (Development Research Center of theState Council of China, 2003). Since almost all of economic
379Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
activities consume energy, energy needs, along with CO2
emissions, will inevitably increase, which will likely createtensions between the country’s need for economic growth,energy and environmental protections. These divergent pres-sures could hinder China’s goal of sustainable development.To address this problem effectively, analysis tools are requiredso as to support energy and environmental policy decisionmaking. For this purpose, this study investigated and devel-oped China’s Energy Requirements & CO2 Emissions Analy-sis System (CErCmA) to assess how changing certain socialand economic policies could impact China’s future energyneeds and CO2 emissions.
Many studies have examined CO2 emissions analysis toolsincluding Silberglitt et al. (2003), Savabi and Stockle (2001),Roca and Alcantara (2001), Gielen and Moriguchi (2002),Clinch et al. (2001), Hsu and Chen (2004), Winiwarter andSchimak (2005), Galeotti and Lanza (2005), Pan (2005),Scrimgeour et al. (2005), Ball et al. (2005) and Sun (1999).Similar investigations into CO2 emissions analysis toolsgeared for China include presenting forecasts of energy con-sumption and related emissions (Lu and Ma, 2004; Chen,2005; Crompton and Wu, 2005; Gielen and Chen, 2001), ana-lyzing strategies for developing a sustainable energy system(Qu, 1992; Wu and Li, 1995; Ni and Thomas, 2004; Xuet al., 2002), assessing impacts of driving forces on historicalemissions (Wu et al., 2005 and Zhang, 2000), exploring vari-ous types of energy technology (Wu et al., 1994; Yan andKong, 1997; Feng et al., 2004; Mu et al., 2004; Eric et al.,2003; Solveig and Wei, 2005), and energy efficiency standards(Lang, 2004 and Yao et al., 2005).
This study aims to extend current studies to obtain not justone scenario but several possible energy requirements andemissions scenarios under different growth paths of variousdriving factors (not just focusing on technology factors butalso focusing on changes in social and economic factors).
The model and software for China’s Energy Requirementsand CO2 Emissions Analysis System (CErCmA) were devel-oped by combining the inputeoutput model (IeO model)with the scenario analysis concept.
China’s energy system is huge and complex with many un-certainties due to the driving forces of energy requirements.The traditional trend extrapolation approach works onlywhen the changes in driving forces follow established paths,but can shed little light on the case of driving forces movingin a brand-new orbit, e.g., certain risks or challenges baffleeconomic development.
The current popular scenario analysis operates in a differentway in that ‘‘it does not try to predict the future but rather toenvision what kind of futures is possible’’ (Silberglitt et al.,2003). Through the description of various possible futurescenarios representing different growth paths, driving forceuncertainties can be taken into account. Under each scenario,the inputeoutput model is employed to assess China’s energyrequirements along with its CO2 emissions.
In recent years much attention has been given to theseissues. Some of these studies perform sensitivity analyses onone or more social and economic factors. Others identify
certain key factors affecting CO2 emissions and evaluate theirimpacts (e.g., Lee and Lin, 2001; Paul and Bhattacharya,2004; Yabe, 2004). Some focus on one factor, such as socio-economic structural change (Kainuma et al., 2000), or con-sumption patterns (Kim, 2002).
Other researchers primarily analyze the impact of certaineconomic activities or governmental policies on energy con-sumption and related CO2 emissions, such as the impact of in-ternational trade (Machado et al., 2001; Sanchez-Choliz andDuarte, 2004; Kondo et al., 1998; Cruz, 2002), and the effectsof certain policy reforms or frameworks (Bach et al., 2002;Christodoulakis et al., 2000).
Our study extends the sensitivity analysis and applies it toChina’s energy and environmental protection issues. Firstly,major energy consumption impact factors are identified: eco-nomic growth, technological changes, population growth,changing consumption and production patterns, and urbaniza-tion. Secondly, we construct a set of future scenarios describ-ing different growth paths based on these factors; then weapply the IeO model to compute energy requirements alongwith CO2 emissions under each scenario.
This paper is organized as follows:
In Section 1 the CErCmA modeling framework isdescribed, along with its underlying rationale, design prin-ciples, model components and a case study.Section 2 presents the rationale for using the IeO model toanalyze China’s energy requirements and CO2 emissionsalong with the system components. Section 3 introducessoftware we developed based on the system explained inSection 2. Section 4 presents an application to assessChina’s energy requirements along with projected CO2
emissions in 2010 and 2020; this is followed by conclusionsand corresponding policy recommendations. Finally,strengths, research challenges and further work needed toimprove the system are presented in Section 5.
2. CErCmA: approach and components
2.1. Basic approach: inputeoutput model
The CErCmA system was established based on the inputeoutput (IeO) model, an analytical framework developed byProfessor Wassily Leontief in the late 1930s. The mainpurpose of the inputeoutput model is to establish a tessellatedinputeoutput table and a system of linear equations.
An inputeoutput table shows monetary interactions orexchanges between the economic sectors and therefore theirinterdependence. The rows of an IO table describe the distri-bution of a sector’s output throughout the economy, whilethe columns describe the inputs required by a particular sectorto produce its output (Miller and Blair, 1985).
The system of linear equations also describes the distribu-tion of a sector’s output throughout the economy mathemati-cally, i.e., sales to processing sectors as inter-inputs or to
380 Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
consumers as final demand. The matrix notation of this systemis:
X ¼ AXþ Y ð1Þ
where (suppose there are n sectors in the economy) X: totaloutput vector with n dimensions whose element Xj is the out-put of sector j; Y: final demand vector with n dimensions (finaldemand consists of household and government consumption,public and private investment, inventory and net-export); A:direct requirement matrix with n� n dimensions, its elementaij denotes the direct requirement of sector j on sector i forper unit output of sector j. A is also called a technology matrix.aij is obtained through:
aij ¼xij
Xj
ði; j ¼ 1;2;.;nÞ ð2Þ
where xij is the monetary value from sector i to sector j.Thus Eq. (1) can be rewritten as:
X ¼ ðI�AÞ�1Y ð3Þ
where I denotes the n� n dimension identity matrix, andðI � AÞ�1 is called the ‘Leontief inverse matrix’ whose ele-ments bijði; j ¼ 1; 2;.; nÞ represent the total amount of com-modity i required both directly and indirectly to produce oneunit of final demand of commodity j.
In this study, the basic IeO model is extended to computeenergy requirements along with CO2 emissions. Only primaryenergy is taken into account to avoid double counting.
2.2. Energy requirements and CO2 emissions model
2.2.1. Energy requirements modelFirstly, the requirements of fossil fuels, i.e., coal, crude oil
and natural gas, are calculated by the following equation:
QF ¼ QPþQR ð4Þ
where QF: total fossil energy requirements, 3� 1 matrix, itselement QF
j represents the requirement of each fossil fuel,i.e., coal, crude oil and natural gas; QP: production energyrequirements, 3� 1 matrix, its element QP
j represents the pro-duction energy requirements of each fossil fuel; QR: householdenergy requirements, 3� 1 matrix, its element QR
j representsthe household energy requirements of each fossil fuel.
QP and QR are calculated, respectively, as follows (Cruz,2002):
QP ¼ CðI�AÞ�1Y ð5Þ
where C: 3� n matrix, its element cij represents the (physical)quantity of fuel i used per unit of total output in sector j.
QR ¼WP ð6Þ
where W: 3� 1 vector, representing per capita householdenergy consumption, each element of which corresponds toone type of fossil fuel; P: population; suppose b is the ratio
of fossil energy over primary energy, then the primary energyrequirements QT can be calculated as:
QT ¼
P3i¼1
QFi
bð7Þ
2.2.2. Energy intensity modelEnergy intensity is the energy consumption per unit of GDP
output, which can be calculated using the following equation:
DQ ¼QT
GDP¼ QTPn
j¼1
Zj
ð8Þ
where DQ: energy intensity; Zj: value added of sector j.
2.3. CO2 emissions model
2.3.1. Total CO2 emissionsAccording to the approach recommended by the IPCC
(PRC Ministry of Science and Technology Economy &Energy e NGO, 2001), the determinant of CO2 emissions dif-fers from that of other greenhouse gases. Since CO2 emissionsrely on the fuel carbon content, the IPCC does not calculateCO2 emissions as it does other greenhouse gases.
The calculation steps are as follows:(1) Introduce a conversion coefficient d to change the unit
of energy consumption from ‘‘oil equivalent’’ to ‘‘106 kJ’’:
d¼ 41:868� 106 kJ=toe
(2) Obtain carbon content multiplied by potential carbonemissions factor matrix.
(3) Correct non-oxidized carbon with the fraction ofoxidized carbon.
Table 1 shows the potential carbon emissions factors of themain fossil fuels and the fraction of oxidized carbon.
(4) Conversion from oxidized carbon to CO2 emissions.Similar to energy requirements, the calculation of CO2
emissions is also divided into industrial production CO2 emis-sions and household CO2 emissions.
M ¼MPþMR ð9Þ
Table 1
Potential carbon emission factors of leading fossil fuels and their fraction of
oxidized carbon
Fuel Potential carbon emissions
factora (kg carbon/106 kJ)
Fraction of
oxidized carbonb
Coal 24.78 0.98
Crude oil 21.47 0.99
Natural gas 15.30 0.995
a Workgroup 3 of the National Coordination Committee on Climate Change
(Xue, 1998).b IPCC (PRC Ministry of Science and Technology Economy & Energy e
NGO, 2001).
381Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
where M: total CO2 emissions; MP: production CO2 emissions;MR: household CO2 emissions; MP and MR can be calculatedby the following equations, respectively (unit: billion tC).
MP ¼ dX3
j¼1
ejhjQPj ð10Þ
MR ¼ dX3
j¼1
ejhjQRj ð11Þ
where ej: the carbon emissions factor of fossil fuel j (unit:kg C/106 kJ); hj: the fraction of oxidized carbon of fuel j.
2.3.2. Other parameters of CO2 emissions
2.3.2.1. Per capita CO2 emissions.
n¼M
Pð12Þ
where v: per capita CO2 emissions.
2.3.2.2. CO2/TPEC. CO2/TPEC describes CO2 emissionsfrom per unit of total primary energy consumption (TPEC).
G¼ M
QT
¼
P3j¼1
dejhjQFj
QT
¼X3
j¼1
dejhj
QFj
QT
ð13Þ
where G: CO2/TPEC.Eq. (13) shows that the change of CO2/TPEC can reflect the
change in energy structure.
2.3.2.3. CO2 emissions intensity.
DM ¼M
GDP¼ M
QT
QT
GDP¼ GDQ ð14Þ
where DM: CO2 emissions intensity.Eq. (14) implies that the trend of the change in CO2 emis-
sions intensity is determined by that of energy intensitybecause energy-related CO2 emissions result from fuel com-bustion, and the variation speed of CO2 emissions intensityis determined by the index of CO2/TPEC (Sun, 2003).
2.4. Combining influences of in driving forces
Eqs. (4)e(8) show that future energy demand, the technol-ogy matrix and the energy efficiency improvement matrixmust first be calculated in order to obtain future energyrequirements. The process to obtain these variables also pres-ents how the changes of driving forces are combined into themodel.
In this paper superscript f indicates that the variable isrelated to terminal year f ; superscript c indicates that thevariable is related to base year c.
2.4.1. Final demand Y f for terminal year fThe procedure to calculate future final demand Yf consists
of three steps:
2.4.1.1. Calculate future per capita expenditure for each sec-tor. Income elasticity measures the percent change of expen-diture induced by a percent change of income. That is:
3¼Kf �Kc
Kc
Lf � Lc
Lc
ð15Þ
where 3: income elasticity; K f: per capita expenditure for ter-minal year f; Kc: per capita expenditure for base year c; Lf: percapita income for terminal year f; Lc: per capita income forbase year c.
Transforming Eq. (15) for the future per capita expenditureis obtained as follows:
Kf ¼�
1þ 3
�Lf � Lc
�Lc
�Kc ð16Þ
Similarly Kfu and Kf
r can be obtained, where Kfu: urban per
capita expenditures for terminal year f; Kfr : rural per capita
expenditures for terminal year f.
2.4.1.2. Calculate future aggregate household consump-tion. The aggregate household consumption can be consideredas the product of per capita expenditures and total population.Urban and rural aggregate household consumption should becalculated separately, since there is a huge gap between cityand countryside consumption patterns and standard of living.
Tf ¼ Tfu þ Tf
r ¼ KfuPf hþKf
r Pf ð1� hÞ ð17Þ
where Tf: aggregate household consumption for terminal yearf; Tf
u and Tfr : urban and rural aggregate household consumption
for terminal year f; h: urbanization rate, i.e., the share of urbanpopulation in the total population of the country, for terminalyear f.
2.4.1.3. Estimating future final demand. Here the future finaldemand Yf: is estimated from the results of future aggregatehousehold consumption.
Yf ¼ Tf
q fð18Þ
where q f denotes the ratio of aggregate household consump-tion over final demand.
2.4.2. Direct requirement matrix Af for terminal year fHere, the RAS approach (Miller and Blair, 1985) is
employed to obtain the future direct requirement matrix.The RAS method is a common tool used to update the
inputeoutput matrix Af. It attempts to estimate the n� n tech-nology coefficients from three types of information for the
382 Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
year of interest. Regarding this study, the information neededis as follows:
(1) future total output for sector i, Xfi ;
(2) future total intermediate deliveries for sector i, Ufi , which
is equal toPn
j¼1xij, and equals sector’s total output Xfi mi-
nus sector final demand Yfi ;
(3) future total sector intermediate purchases for sector i, Vfi ,
which is equal toPn
i¼1xij, and equals Xfi minus sector
value added Zfi .
The basic aim of RAS is that according to the substitutionand manufacturing assumption about technology change, Xf
i ,Uf
i , Vfi are used to obtain row multipliers (R) and column mul-
tipliers (S), which are used to modify each row and each col-umn of the base year direct requirement matrix Ac,respectively. These two multipliers can be obtained by usingan iterative algorithm, illustrated in Fig. 1. Next, the futuredirect requirement matrix Af can be calculated as follows:
Af ¼ RAcS: ð19Þ
2.4.3. Cf and Wf
The future energy use, per unit of output, matrix C f, and percapita household energy consumption matrix W f need to be ob-tained to explicitly embody the impacts of changes in technol-ogy and energy/environment policy on energy requirementand related CO2 emissions. The improvements from Cc andW c to C f and W f can be obtained by referring to current energylayouts.
3. Software
Software to forecast future energy requirements and energyintensity as well as CO2 emissions using the model above hasbeen developed using Visual Basic 6.0, called CErCmA 1.0.
3.1. Characterization of main components
As illustrated in Fig. 2, CErCmA is composed of four parts:
� Database and database management subsystem: to input,store and manage base year data, factor scenarios andmodeling results.� User interface: to provide graphic interface so as to conve-
niently input parameters and search results.� Scenario construction subsystem: to combine factor sce-
narios into the model and generate integrated scenarios.� Model base subsystem: to simulate the situation in termi-
nal years for each constructed scenario.
As its result, the system generates the following: energyrequirements for each sector and for each fuel type, energy in-tensity for each sector, CO2 emissions for each sector and eachfuel type, CO2 intensity for each sector, CO2 emissions fromper unit of the total primary energy consumption (CO2
emissions/TPEC), per capita energy requirements and CO2
emissions.
3.2. Database and database management subsystem
3.2.1. DatabaseThe database in this system consists of the base year data-
base, the factor scenario database and the result database.The base year database contains base year inputeoutput
tables, socio-economic data and technology data.The factor scenario database includes an economic scenario
base, a population scenario base, an urbanization scenario baseand a technology scenario base. Data in each scenario basecome from the latest statistics released by the governmentand other authorities listed in the references.
� The economy scenario base contains data on the GDPgrowth rate, per capita income growth rate, variations ofincome elasticity and industrial structure.
f
jjV V ?
END
Y
Y
N
N
Scenario construction subsystem
f
iY
f
iZ
Ac
f
iX
f
iV
f
iU
c
ij
f
ijaa
n
j
f
j
f
ijiXaU
1
n
i
f
j
f
ijiXaV
1
i
f
i
i
U
U
r
f
iji
f
ijara
j
f
j
j
V
V
s
j
f
ij
f
ijsaa
f
iiU U ?
Base yeardatabase
Fig. 1. Iterative algorithm of RAS.
383Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
Scen
ario
co
nstru
ctio
n
su
bsystem
Da
ta
ba
se
&
d
ata
ba
se
ma
na
ge
me
nt s
ub
sy
ste
m
Factor scenario database
Base year database Resultsdatabase
Final demand estimation
Technology change estimation
Mo
del b
ase su
bsystem
Energy requirement model
CO2 emissions model
User in
terface
User
Future per capita income Future income elasticity Future population Future urbanization
GDP growth Industrial structure Energy layouts
per capita income populationurbanization aggregate householdconsumption
Direct requirement matrix Energy use per unit outputmatrixPer capita household energyconsumption matrix
Future finaldemand
Energyreuirements
Total CO2emissions
Per capita CO2emissions
CO2 /TPEC
CO2 emission intensities
Future direct requirement matrix Future energy use per unit outputmatrixFuture per capita householdenergy consumption matrix
NOTES: DatabaseCalculation components
Data and data flow
Main output
Energy intensities
Fig. 2. Software structure of CErCmA 1.0.
� The population scenario base contains data on yearly totalpopulation.� The urbanization scenario base contains data of the yearly
urbanization rate.� The technology scenario base contains data on rates of
change in the energy consumption per unit output, rateof change in per capita household energy consumption,etc.� The results database contains the scenario analysis results
including energy requirements and related CO2 emissionsplus additional details.
3.2.2. Database management subsystemThe database management subsystem provides convenient
editing functions that allow users to add, modify and displaydata to both the factor scenario base and results base.
Users can add custom factor scenarios to the factorscenario base when running the system. But without adminis-trators’ authority to confirm these operations, the customscenarios are automatically deleted when the program shutsdown so as to maintain the validity and consistency of thesystem.
Users can create tables and print results for the results base.
384 Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
3.3. Software description
The main interface consists of two parts. The right side ofthe main interface is the welcome interface, along with logininformation for users. If a user logs in as an administrator,the interface will display the corresponding username andrights of the user. Beneath the welcome interface is a fast-entrance for administrators.
The left side of the main interface displays the scenario setsof each driving factor. Users can set scenarios for each factorand click the ‘‘scenario generation’’ icon on the toolbar to ac-tivate them; next, a synthesis scenario appears and is demon-strated on the right side. When the user constructs all thescenarios of concern, users can click the ‘‘scenario analysis’’icon and the analysis results will be generated and automati-cally transferred and inserted into tables and graphics forimmediate presentation.
4. An application: China’s energy requirements andrelated CO2 emissions for 2010 and 2020
In this application the analysis years are set at 2010 and2020.
In this scenario 2010 is the terminal year of the 11thFive-year plan; 2020 is when the Chinese aims to realize itsgoal of building an economically secure society. The govern-ment (Development Research Center of the State Council,2003) set the following objectives: ‘‘On the basis of optimizedstructure and better economic returns, efforts will be made toquadruple the GDP of 2000 by 2020’’ and ‘‘achieve industri-alization by 2020.’’ During this period, major changes are ex-pected in the economy, population, urbanization andtechnology, which will further increase both production andhousehold energy demand. The main question here is: Howwill changes in major social and economic factors impactChina’s future energy requirements and CO2 emissions? CouldChina maintain its advantage of low per capita CO2 emis-sions? Therefore, it is of great significance to assess the energyrequirement and related CO2 emissions in these two years.
4.1. Model assumption
Since the latest inputeoutput tables available for China arefrom 1997, in this application the base year is set to 1997.
Six production sectors and a residential sector are consid-ered in this application: agriculture, manufacturing, construc-tion, transportation, commerce and service, as well asresidential energy use. Primary energy is divided into fourgroups: coal, oil, natural gas, and hydro and nuclear power.
4.2. Scenario
The following five scenarios are established around varioussocial and economic factors affecting energy requirements; thescenarios represent five distinct growth paths that China mightfollow in the future (Table 2).
4.2.1. Economy scenariosThe Development Research Center of the State Council
(2003), led by researcher S. Li, identifies two types ofeconomic development as shown in Table 3. This study alsopredicts the future industry structure, including the percent-ages for primary, secondary and tertiary industries, whichare, respectively, 10.6:54.2:35.2 for year 2010, and7.0:52.6:40.4 for year 2020.
4.2.2. Population scenariosOur study utilizes the forecasts of the Quantitative
Economics Institute of the Chinese Academy of SocialSciences (CASS) and UNEP (Zhou, 2000) (Table 4).
4.2.3. Urbanization scenariosOur study refers to the findings of the Institute of
Geographic Sciences and Natural Resources Research of theChinese Academy of Sciences (Liu et al., 2003). This studyexamines three scenarios:
- High scenario: China’s market-oriented reforms will bea complete success and significantly hasten the urbaniza-tion process. The urbanization rate will be 44.7% in2010 and 54.7% in 2020.
- Medium scenario: China’s market-oriented reforms will bea partial success and the urbanization trend will follow thecommon S-curve trajectory, experienced in the past by
Table 2
Scenario description
Scenarios Scenario description
Scenario A1
(low economy)
Various challenges and risks constrain economic
development and urbanization advancement, and
cause technology to advance at a lower speed
than in scenario B.
Scenario A2
(low economy
þ technology)
The economy and urbanization scenarios are the
same as those in scenario A1, while supposing
that the government initiative efforts to maintain
a technological improvement to achieve its
preset objects of a sound national energy plan.
Scenario B
(business as usual)
Assume in the coming 20 years China’s economy
could maintain its current growth rate and realizes
a relatively high growth rate; per capita income
achieves the preset ‘well-off’, ‘developed’ society
objectives; population and urbanization rate grow
at a medium speed; technology improvement
meet the preset objectives of the PRC’s national
energy plan.
Scenario C1
(Bþ high population)
On the base of scenario B, assume population
growth at a much higher speed resulting in a
new population peak.
Scenario C2
(Bþ technology)
On the basis of scenario B, assume technology
advances at a higher rate than in scenario B.
Table 3
Forecast of economic growth (%)
Scenarios Year 2001e2010 Year 2011e2020
Economy-base 7.9 6.6
Economy-low 6.6 4.7
385Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
most countries. In this scenario the urbanization rate willbe 43.03% in 2010 and 50.14% in 2020.
- Low scenario: China’s market-oriented reforms will notsignificantly advance economic progress and the urbaniza-tion process will still be constrained by the system as inthe past 20 years. The urbanization rate will be 42.24%in 2010 and 48.25% in 2020.
4.2.4. Technology scenarios
4.2.4.1. Technology improvement matrix. Table 5 presents thethree sets of technology improvement scenarios.
4.2.4.2. Ratio of renewable energy over primary energy. Thisstudy utilizes the forecasts of the Academy of MacroeconomicResearch Workgroup of the State Development PlanningCommission (1999a) (Table 6).
4.3. Data
Table 7 describes the data sources for the study.
4.4. Model checking
With the data from 1997, we checked our model for theperiod 1998e2003, as shown in Fig. 3. Generally, the simula-tion results are close to the actual statistical data, with thelargest relative error 11.19%, the smallest relative error0.90%, and the average relative error 2.21%.
4.5. Simulation results and policy implication
In this section, the results of energy requirements and re-lated CO2 emissions in each scenario for 2010 and 2020 arediscussed.
Table 4
Forecast of China’s population (in billions)
Scenarios (organizations) 2010 2020
High (CASS) 1.48 1.52
Low (UNEP) 1.39 1.45
Table 5
Technology scenarios
Scenario Description
Low Energy efficiency achieves a 5% less improvement than
the medium case.
Medium The improvement of energy efficiency achieves the goal
of the National Energy-Saving Layout (Academy of
Macroeconomic Research, State Development Planning
Commission, 1999b, see Appendix A for a brief introduction).
High Energy efficiency achieves a 5% greater improvement than
the medium case.
4.5.1. Total and per capita energy requirements andrelated CO2 emissions
Fig. 4 presents the results of energy requirements alongwith CO2 emissions in the model scenarios. In 2010, the totalenergy requirement will be 1.57e1.84 billion tons of oilequivalent (toe), and the corresponding CO2 emissions willbe 1.33e1.57 billon tC; in 2020, the total energy requirementwill be 1.88e2.64 billion toe, and corresponding CO2 emis-sions will be 1.54e2.17 billion tC.
An analysis of the forecast results for these scenarios showsthe followings.
4.5.1.1. Energy efficiency plays an important role in the controlof energy consumption and related CO2 emissions. ScenarioA1 is one of the two scenarios with the highest level of energyneeds and CO2 emissions. According to the scenario construc-tion, economic growth and urbanization advancement in thisscenario are the lowest of all, which means the main driverof energy consumption for final demand is the lowest. Butthe improvement speed of energy efficiency in this scenariois also the lowest; and thus the terminal energy requirementand related CO2 emissions in this scenario are higher thanthose in other scenarios.
On the other hand, economic growth and urbanization ad-vancement in scenario C2 occur at the same speed as in sce-narios B and C1, and at a higher speed than in scenarios A1and A2. Since improvement speed of energy efficiency is thehighest in this scenario, the terminal energy requirement andrelated CO2 emissions in this scenario are lower than in otherscenarios except scenario A2.
Through this analysis it is clear that energy efficiencyimprovement plays an important role in the control of energy
Table 6
Forecast of ratio of renewable energy over primary energy (%)
Ratios 2010 2020
Hydro power 9.3 10.3
Nuclear power 2.3 3.64
Table 7
Data sources
Variable
(matrix)
Source
Ac, Yc, Tcu,
Tcr , qf a
Inputeoutput table of China (Department of National Accounts,
National Bureau of Statistics, P.R. China, 1997)
3fu, 3f
r A modification of Hubacek and Sun (2001),
see Appendix B
Lcu, Lc
r , Pc, hc China Statistical Yearbook (National Bureau of Statistics, P.R.
China, 1998)
Cc, Wc Combining the data in energy balance tables
1997 with the corresponding data in
inputeoutput table of China (Department of National Accounts,
National Bureau of Statistics, P.R. China, 1977; Department of
Industrial and Transportation Statistics, National Bureau of
Statistics, P.R. China, 2001)
a The ratio of the construction sector is zero in the yearbook of 1997. Here
this ratio is set to be 15% which is allocated for urban and rural household
consumption based on the ratio between urban and rural population (Hubacek
and Sun, 2001).
386 Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
500.00
600.00
700.00
800.00
900.00
1000.00
1100.00
1200.00
simulation results statistic data
simulation results 924.56 813.91 881.53 893.26 955.89 1106.54 1138.06 statistic data 924.56 887.72 873.66 874.85 905.85 995.20 1126.66
1997 1998 1999 2000 2001 2002 2003
Fig. 3. Energy demand (requirements) for 1998e2003 (million toe).
consumption and related CO2 emissions. Great efforts shouldbe taken to improve energy efficiency in final demand sectors,to enhance energy conversion efficiency, and to graduallycreate a lower energy consumed product system and a lifesystem.
4.5.1.2. Population has a significant impact on energy require-ments along with CO2 emissions. Scenario C1 is the ‘‘highpopulation’’ scenario in which the energy requirements and re-lated CO2 emissions will be second place in 2010, and rise tofirst place in 2020. So even with the development of technol-ogy, energy requirements along with CO2 emissions will stillincrease rapidly if planners and policymakers fail to controlpopulation at the same time.
4.5.1.3. Energy efficiency is highly related to economicgrowth. Energy efficiency in scenario A2 improves slowerthan that in scenario C2, but economic growth and urbaniza-tion rates are also slower in this scenario than in scenario
C2. The result is that energy consumption along with CO2
emissions in this scenario for the two final years are both lowerthan in scenario C2. It is obvious that the low energy con-sumption and low CO2 emissions in scenario A2 is obtainedat the cost of a low economic growth rate.
The above results show that total energy requirements andCO2 emissions increase rapidly in all scenarios. As for percapita results, Fig. 5 compares per capita CO2 emissions ineach scenario with the world average in 2003. It appears thatin 2010, per capita CO2 emissions in scenario A1 are slightlyhigher than the world average in 2003. Per capita CO2 emis-sions in all the scenarios will rise until 2020 by more thanthe world rate in 2003.
4.5.2. Energy structure and CO2 emissions by energy typeThe forecast of energy structure and CO2 emissions by
energy type are shown in Figs. 6 and 7, respectively.Variations of energy structures in all scenarios are similar in
general. The proportion of coal decreases, but still occupies
0.00
0.50
1.00
1.50
2.00
2.50
3.00
A1 0.94 0.94 0.94
0.94 0.94
1.84 2.50 0.87 1.57 2.06 A2 1.57 1.88 0.87 1.33 1.54 B 1.67 2.52 0.87 1.42 2.08
C1 1.78 2.64 0.87 1.51 2.17 C2 1.59 2.31 0.87 1.35 1.90
1997 2010 2020 1997 2010 2020
Energy requirements (billion toe) CO2 emissions (billion tC)
Fig. 4. Results of energy requirements and related CO2 emissions in assigned scenarios.
387Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
0.50.6
0.70.8
0.91
1.11.2
1.31.4
1.5
1997 2010 2020
ScenarioA1 ScenarioA2 ScenarioB
ScenarioC1 ScenarioC2 World Average 2003
Per capita CO2 emissions (tC/capita)
Fig. 5. China’s per capita CO2 emissions in 2010 and 2020 vs. world average level in 2003.
the principal usage; the proportions of oil and natural gas rises,but the share of natural gas use is still low.
As shown in Table 1, the potential emissions factor of coalis much larger than that of oil and natural gas. Thus, generally,the energy structure tends to impact more significant on CO2
emissions control. But since coal still dominates the energystructure, CO2 emissions by coal is the largest emissions factorin the future two years (see Fig. 7). So it appears that thepotential to control CO2 emissions by adjusting the energystructure during this period is limited.
4.5.3. Energy intensity, CO2 intensity and CO2/TPECIn 2010, energy intensity will be 0.79e1.023 toe/104 Yuan,
CO2 intensity will be 0.671e0.87 tC/104 Yuan, and CO2/TPEC will be about 0.85 tC/toe. In 2020, energy intensitywill be 0.604e0.864 toe/104 Yuan, CO2 intensity will be0.497e0.711 tC/104 Yuan, and CO2/TPEC will be about0.823 tC/toe.
Table 8 shows the change rate of energy intensity and CO2
intensity in 2010 vs. 1997, and 2020 vs. 2010.
0% 20% 40% 60% 80% 100%
Base year
2010 A1
2010 A2
2010 B
2010 C1
2010 C2
2020 A1
2020 A2
2020 B
2020 C1
2020 C2
Coal Oil Natural Gas Other
Fig. 6. Energy structure in constructed scenarios (%).
In the two final years, both the lowest energy intensity andCO2 intensity appear in scenario C2, and both the highest en-ergy intensity and CO2 intensity appear in scenario A1.
In general, energy intensity and CO2 intensity are declin-ing. CO2 intensity is declining faster than energy intensity be-cause CO2/TPEC, another factor determining CO2 intensity, isalso declining.
But in all the scenarios the decline of CO2/TPEC is slow.This implies that in two final years the coal-dominated energystructure is changing very little. CO2/TPEC determines thevariation speed of CO2 intensity. The slow declining speedof CO2/TPEC to a great extent limits the decline of CO2 emis-sions in this period.
4.5.4. Relationship between CO2 emissionsand driving factors
According the definition of CO2 emissions, the followingdecomposition was performed to assess the impacts of majordriving forces on CO2 emissions:
CO2 emissions ¼ CO2 intensity�GDP
¼ CO2=TPEC� energy intensity�GDP
0% 20% 40% 60% 80% 100%
Base year
2010 A1
2010 A2
2010 B
2010 C1
2010 C2
2020 A1
2020 A2
2020 B
2020 C1
2020 C2
Coal Oil Natural Gas
Fig. 7. CO2 emissions by energy type in constructed scenarios (%).
388 Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
¼ CO2=TFEC
� the ratio of fossil energy over primary energy
� energy intensity� per capita GDP� population
where, CO2/TFEC presents CO2 emissions from the total fos-sil energy consumption.
The variation of CO2 emissions and that of each drivingforce are calculated, respectively, as presented in Fig. 8.
The decomposition implies that, in the absence of extraenergy or environmental policies, among all the drivingfactors, per capita GDP plays the most important role, energyintensity comes in second, and population, third, while the
Table 8
Change rate of energy intensity and CO2 intensity
Scenarios 2010 vs. 1997 2020 vs. 2010
Energy intensity CO2 intensity Energy intensity CO2 intensity
A1 0.82 0.76 0.84 0.82
A2 0.70 0.65 0.75 0.72
B 0.67 0.62 0.79 0.77
C1 0.71 0.66 0.78 0.75
C2 0.64 0.59 0.76 0.74
0 0.5 1 1.5 2 2.5 3
A1
A2
B
C1
C2
A1
A2
B
C1
C2
populationper capita GDPenergy intensityratio of fossil energy over primary energyCO2/TFES
CO2 emissions
2010-2020
1997-2010
Fig. 8. Relationships between CO2 emissions and driving factors.
driving forces of the ratio of fossil energy over primary energyand that of CO2/TFES are quite small.
In all the scenarios, the forward driving effect of per capitaGDP growth is much larger than the backward driving effectsof other factors, which can explain why CO2 emissions aregrowing so fast.
Between 2010 and 2020, the backward driving effect of en-ergy intensity is smaller than between 1997 and 2010, whichshows that further technology enhancements will becomemore difficult. Therefore, on the one hand, certain policiesshould be devised to promote changes in energy structureso as to effectively make use of the backward driving effectof the ratio of fossil energy over primary energy and that ofCO2/TFES. On the other hand, changing energy structurewould be a long-term task, the effect of which is not quitelikely to be seen before 2020 (Zhou and Yukio, 1996), sothe coal-dominated energy structure is not expected to changetoo much in the near future, the decline of energy intensity islimited; correspondingly the decline of CO2 intensity is alsolimited. Moreover, during this period China’s economy willstill grow at a relatively high speed, so the potential for CO2
mitigation is limited.
4.5.5. Sector energy requirements and CO2 emissionsFigs. 9 and 10 show the results of sector and residential en-
ergy requirements and CO2 emissions, respectively.Variations in all the scenarios are similar. The energy need
in manufacturing sector is declining, which shows the impactof the accelerated development of service and transportationcaused by accelerated urbanization. But due to the govern-ment’s objective ‘‘to achieve industrialization by 2020,’’ thescale of manufacturing will still expand. The energy require-ments and CO2 emissions of this sector will still take the larg-est share, followed by the transportation sector.
The share of energy for the agriculture sector first increasesand then declines: the share in 2010 is larger than in 1997,while the share in 2020 is smaller than in 2010, but is stilllarger than in 1997. This phenomenon can be explained bythe two driving forces of opposite directions on agriculture en-ergy consumption, i.e., the continuous decrease in agriculturalenergy usage in the industrial structure has a backward impact,while the continuous increase in the agricultural mechaniza-tion level has a positive impact.
The shares of construction and commerce in total energyconsumption increase rapidly. The shares of transportationand the service industry in 2010 rise markedly from the1997 level, while their shares in 2020 vary not much fromthose in 2010, they rise just a little.
4.5.6. Sector energy intensity and CO2 intensityTables 9 and 10 present the results of sector energy inten-
sity and CO2 intensity, respectively.In the two final years, the energy and CO2 intensities of the
manufacturing and transportation sectors are evidently higherthan average.
For manufacturing, energy and CO2 intensities in scenarioA1 are much higher than those in the other four scenarios.
389Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
0%
60%
70%
80%
90%
100%
Baseyear
2010 A1 2020 A1 2010 A2 2020 A2 2010 B 2020 B 2010 C1 2020 C1 2010 C2 2020 C2
Agriculture Manufacturing Construction TransportationCommerce Service Residential
10%
20%
30%
40%
50%
Fig. 9. Sector energy requirement in each scenario (%).
For transportation, in 2010, energy intensity in all scenarios isquite close to that of manufacturing.
Manufacturing and transportation take the largest shares ofenergy and produce the largest amount of CO2 emissions, thusthe energy and CO2 intensities of these two sectors will toa great extent impact the speed of energy requirements andCO2 emissions growth. These two sectors should be pivotalin improving energy efficiency.
Among all scenarios, scenario A1 is the worst path in termsof controlling the fast growth trends of energy requirementsand CO2 emissions.
Energy and CO2 intensities in agriculture, construction,commerce and the non-material sectors remain lower thanthe sector average. The intensities of the construction sectorare the lowest in the two final years, but tend to rise rapidly.As shown above, the shares of construction and commercein total energy consumption increase rapidly, there could bean induced reduction in CO2 emissions. Therefore great effortsshould be taken to maintain the low energy and CO2 intensitiesin these sectors.
4.6. Policy implications
Summarizing the above forecast results and analyses of themodel, the following policy recommendations are proposed.
4.6.1. Policies promoting changes in energy structureshould be devised to control the quickly increasingCO2 emissions
The above forecast results show that even in scenario C2,where technology advances the fastest, CO2 emissionsincrease rapidly. Moreover, the increase of per capita emis-sions will quite possibly reach or exceed the worldwide aver-age within 20 years. The results also show that with timegoes on further enhancements in energy-saving technologywill become more and more difficult. Therefore, it is desir-able to establish energy and environmental policies in favorof cleaner energies as early as possible to accelerate thechanges in energy structure and to support sustainable eco-nomic development.
60%
80%
100%
AgricultureManufacturing
ConstructionTransportation
CommerceService
Residential
0%
20%
40%
Baseyear
2010 A1 2020 A1 2010 A2 2020 A2 2010 B 2020 B 2010 C1 2020 C1 2010 C2 2020 C2
Fig. 10. Sector CO2 emissions in each scenario (%).
390 Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
4.6.2. Effective policies should be developed andimplemented to encourage governmental agenciesand corporations to increase energy efficiency
The model results clearly show that in all scenarios energyintensity has an obvious backward driving effect on CO2 emis-sions, second only to the forward driving effect of per capitaGDP. What’s more, efforts to change China’s coal-dominatedenergy structure will take a long term to take effect, thus en-ergy efficiency improvement will still play a pivotal role inCO2 mitigation.
The manufacturing and transportation sectors consume themajor share of energy in China, and their energy intensities arevery high, much higher than the sector average. Model resultsshow that their share of emissions is decreasing but will main-tain the main part of total CO2 emissions in recent 20 years.The high CO2 intensity of manufacturing is determined byits high energy intensity. Therefore, certain policies must becontinuously implemented to decrease the manufacturing sec-tor’s energy intensity, such as accelerating the adjustment ofthe industrial and product structures within the manufacturingsectors to improve manufacturing energy efficiency.
The shares of construction and commerce in total energyconsumption increase rapidly, thus there is an induced reductionin CO2 emissions because of the low CO2 intensities of thesetwo sectors. Therefore particular attentions should also bepaid to the energy efficiency improvements in these two sectors.
4.6.3. The potential for CO2 mitigation in China is limited inthe next 20 years, and thus decision making should bebased on this key point
From today until the 2020, China’s GDP is expected tomaintain a high growth rate. While the special coal-dominatedenergy structure is not expected to change much in the nearfuture, and because the efficiency of coal is much lower
Table 9
Energy intensity for each sector (toe/104 Yuan)
Sector 2010 2020
A1 A2 BAU C1 C2 A1 A2 BAU C1 C2
Agriculture 0.35 0.31 0.29 0.31 0.28 0.35 0.27 0.27 0.28 0.25
Manufacturing 1.86 1.58 1.52 1.61 1.44 1.47 1.11 1.17 1.23 1.07
Construction 0.18 0.15 0.14 0.15 0.13 0.23 0.16 0.17 0.17 0.15
Transportation 1.95 1.55 1.49 1.58 1.41 1.88 1.26 1.29 1.35 1.17
Commerce 0.30 0.26 0.25 0.27 0.24 0.26 0.21 0.21 0.22 0.20
Service 0.29 0.26 0.25 0.26 0.24 0.26 0.21 0.21 0.22 0.20
Sector average 1.02 0.87 0.83 0.89 0.79 0.86 0.65 0.66 0.69 0.60
Table 10
Sector CO2 intensity (tC/104Yuan)
Sector 2010 2020
A1 A2 BAU C1 C2 A1 A2 BAU C1 C2
Agriculture 0.29 0.25 0.24 0.25 0.23 0.27 0.21 0.21 0.22 0.20
Manufacturing 1.61 1.37 1.32 1.40 1.25 1.24 0.94 0.99 1.04 0.91
Construction 0.14 0.12 0.11 0.12 0.11 0.18 0.13 0.13 0.14 0.12
Transportation 1.50 1.20 1.15 1.22 1.09 1.41 0.95 0.97 1.01 0.88
Commerce 0.25 0.21 0.21 0.22 0.20 0.21 0.17 0.17 0.18 0.16
Service 0.22 0.20 0.19 0.20 0.18 0.20 0.16 0.16 0.17 0.15
Sector average 0.87 0.74 0.71 0.75 0.67 0.71 0.53 0.54 0.57 0.50
than that of oil and natural gas in most cases (He et al.,1995), the decline of energy intensity will likely be limitedover the next 15 years. At the same time, the high cardinalnumber of coal, to a great extent, will likely limit the decreaseof CO2 emissions by total fossil fuel consumption, which canstill cover the majority of total primary energy consumption.Consequently CO2 intensity will likely not drop considerablyfast in the next 20 years.
5. Discussion and perspectives
The application shows that CErCmA provides an effectivetool for assessing energy requirements along with CO2 emis-sions in China. The combination of scenario analysis and theinputeoutput model provides not only a thorough, integratedanalysis, but also a means to explicitly assess the impact ofeach driving force.
Further work is needed to address the followings in order toimprove our model.
5.1. The limitation of the static IeO model
So far, the methodology of the dynamic IeO model isstill being developed. For our current system we chose theclassical static IeO model. However, a complete staticapproach would fail to identify the structural changes. Toaddress this problem, we applied an adjustment to the baseyear direct requirement matrix Ac by using the RAS approachto obtain a possible structural change in this paper. In thisway, we could simulate the structural changes during thetime horizon of our study.
Nevertheless, the RAS approach has some weakness,primarily its economic assumptions: the sector consistencyof substitution and manufacturing impacts is not satisfied inmany cases. One way to improve the RAS approach may beto combine it with a case study of the key coefficients in a di-rect requirement matrix. Today, a great deal of attention is be-ing focused on improving this approach. Accordingly, in thefuture, we plan to trace the development of the RAS approachand other possibly more effective adjustment approaches toimprove the precision of the current system.
5.2. Technology scenario base
Because of data availability limitations, the technologyscenarios in the current version were attained by adjusting cur-rent energy layouts. This kind of data source would, to someextent, weaken the plausibility of the technology scenariosand limit the assessment of the impact of technology in termsof depth. In future studies, researchers should try to improvethe technology scenario base.
5.3. Carbon tax analysis: quantifying policyrecommendations
Quantifying the feedback of certain policies couldenable us to test the possible executive effects of the
391Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
policies. Carrying out a carbon tax analysis can be a possibleapplication if the carbon tax can be incorporated into themodel.
5.4. Models on the regional level
Current version only includes models on the nationallevel. However, social and economic factors, such aspopulation intensity and income levels, as well as energy effi-ciency significantly differ among regions in China. Modeling en-ergy requirements and CO2 emissions on a regional level wouldbe of great help to energy and environment policy making.
Acknowledgements
The authors gratefully acknowledge the financial supportfrom the National Natural Science Foundation of China(NSFC) under the grants Nos.70425001, 70573104 and70371064, the Key Projects from the Ministry of Science andTechnology of China (grants 2001-BA608B-15, 2001-BA605-01), Harvard University Committee on Environment China Pro-ject. Ying Fan would like to thank Prof. Tom Lyons, Prof. Yong-miao Hong and Ms. Deborah Campbell at Cornell University fortheir valuable comments and kind help. Yi-Ming Wei truly ap-preciates the supports from Prof. Michael B. McElroy and Mr.Chris P. Nielsen at Harvard University. We also would like tothank Prof. A. J. Jakeman and the other four anonymous refereesfor their helpful comments on the earlier draft of our paper ac-cording to which we improved the content.
Appendix A. A brief introduction of the NationalEnergy-Saving Layout
The Group of the Academy of Macroeconomic Research ofthe State Development Planning Commission figured out the po-tential of energy efficiency improvement in sectoral final energyconsumption through adopting various techniques and mea-sures. (Workgroup of the Academy of Macroeconomic Re-search of the State Development Planning Commission, 1999b).
(1) Iron and steel industry.
Energy efficiency can rise by 12%, from 46% in 1995 to58% in 2010, a bit higher than that of industrialized coun-try in early 1970s.
(2) Materials industry.
Energy use per unit output will decrease. Annually energysave rate will be 1.5%. Energy efficiency will increase by10%, from 40% in 1995 to 50% in 2010.
(3) Building materials industry.
Energy use per unit output will decrease. Annuallyenergy-saving rate will be 1.4%. Energy efficiency
will increase by 10.5%, from 45% in 1995 to 55.5%in 2010.
(4) Transportation industry.
Energy efficiency in 2010 will be a little higher than thatin 1995.
(5) Household energy consumption.
In 2010, urban household energy efficiency is hopefully toreach 50% of the level of developed countries in early1990s; rural household efficiency will rise from 25% in1995 to 45%.
(6) Other industries.
In 2010, energy efficiency is hopefully to reach the levelof developed countries in early 1990s, i.e., rise 5 to 10%from the level of 1995.
References
Bach, S., Kohlhaas, M., Meyer, B., Praetorius, B., Welsch, H., 2002. The
effects of environmental fiscal reform in Germany: a simulation study.
Energy Policy 30, 803e811.
Ball, M., Calaminus, B., Kerdoncuff, P., Rentz, O., 2005. Techno-economic da-
tabases in the context of integrated assessment modelling. Environmental
Modelling & Software 20, 1189e1193.
Chen, W.Y., 2005. The costs of mitigating carbon emissions in China:
findings from China MARKAL-MACRO modeling. Energy Policy 33,
885e896.
Christodoulakis, N.M., Kalyvitis, S.C., Lalas, D.P., Pesmajoglou, S., 2000.
Forecasting energy consumption and energy related CO2 emissions in
Greece: an evaluation of the consequences of the Community Support
Framework II and natural gas penetration. Energy Economics 22,
395e422.
Clinch, J.P., Healy, J.D., King, C., 2001. Modelling improvements in
domestic energy efficiency. Environmental Modelling & Software 16,
87e106.
Crompton, P., Wu, Y.R., 2005. Energy consumption in China: past trends and
future directions. Energy Economics 27, 195e208.
Cruz, Luis M.G., 2002. Energyeenvironmenteeconomy interactions: an
inputeoutput approach applied to the Portuguese Case. Paper for the 7th
Biennial Conference of the International Society for Ecological Econom-
ics, ‘‘Environment and Development: Globalisation & the Challenges
for Local & International Governance, ’’ Sousse (Tunisia), 6e9 March
2002.
Appendix B
Income elasticity of various sectors
Sectors 1992e2005 2005e2025
Rural Urban Rural Urban
Agriculture 0.561 0.767 0.509 0.743
Manufacturing 1.100 1.100 1.100 1.100
Construction 1.100 1.100 1.100 1.100
Transportation 1.200 1.200 1.200 1.200
Commerce 1.200 1.200 1.200 1.200
Services 1.200 1.200 1.200 1.200
A modification of Hubacek and Sun (2001).
392 Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
Department of Industrial and Transportation Statistics, National Bureau of
Statistics, P.R. China, 2001. China Energy Statistical Yearbook (1997e
1999). China Statistics Press, Beijing (in Chinese).
Department of National Accounts, National Bureau of Statistics, P.R. China,
1997. InputeOutput Table of China. China Statistics Press, Beijing. (in
Chinese).
Development Research Center of the State Council, 2003. China’s economy
will continue developing quickly in the next 20 years. Research Report
by the Development Research Center of the State Council, China (in
Chinese).
Eric, D.L., Wu, Z.X., Pat, D., Chen, W.Y., Gao, P.F., 2003. Future impli-
cations of China’s energy-technology choices. Energy Policy 31,
1189e1204.
Feng, W., Wang, S.J., Ni, W.D., Chen, C.H., 2004. The future of hydrogen
infrastructure for fuel cell vehicles in China and a case of application in
Beijing. International Journal of Hydrogen Energy 29, 355e367.
Galeotti, M., Lanza, A., 2005. Desperately seeking environmental Kuznets.
Environmental Modelling & Software 20, 1379e1388.
Gielen, D., Chen, C.H., 2001. The CO2 emission reduction benefits of Chinese
energy policies and environmental policies: a case study of Shanghai,
1995e2020. Ecological Economics 39, 257e270.
Gielen, D., Moriguchi, Y., 2002. Modelling CO2 policies for the Japanese
iron and steel industry. Environmental Modelling & Software 17,
481e495.
He, J.K., Zhang, A.L., Lu, L.X., 1995. The technical option of CO2 emissions
mitigation in future energy systems. World Environment 2, 23e27 (in
Chinese).
Hsu, C.C., Chen, C.Y., 2004. Investigating strategies to reduce CO2 emissions
from the power sector of Taiwan. International Journal of Electrical Power
& Energy Systems 26, 487e492.
Hubacek, K., Sun, L.X., 2001. A scenario analysis of China’s land use
and land cover change: incorporating biophysical information into
inputeoutput modeling. Structural Change and Economic Dynamics 12,
367e397.
International Energy Agency, 2003. World Energy Statistics.
Kainuma, M., Matsuoka, Y., Morita, T., 2000. Estimation of embodied CO2
emissions by the general equilibrium model. European Journal of Opera-
tional Research 122, 392e404.
Kim, J.H., 2002. Changes in consumption patterns and environmental degrada-
tion in Korea. Structural Change and Economic Dynamics 13, 1e48.
Kondo, Y., Moriguchi, Y., Shimizu, H., 1998. CO2 emissions in Japan:
influences of imports and exports. Applied Energy 59, 163e174.
Lang, S.W., 2004. Progress in energy-efficiency standards for residential
buildings in China. Energy and Building 36, 1191e1196.
Lee, C.F., Lin, S.J., 2001. Structural decomposition of CO2 emissions from
Taiwan’s petrochemical industries. Energy Policy 29, 237e244.
Liu, S.H., Li, X.B., Zhang, M., 2003. Scenario Analysis on Urbanization and
RuraleUrban Migration in China. Institute of Geographic Sciences,
Beijing, China. Interim Report, IR-03e036.
Lu, W., Ma, Y., 2004. Image of energy consumption of well-off societies in
China. Energy Conversion and Management 45, 1357e1367.
Machado, G., Schaeffer, R., Worrell, E., 2001. Energy and carbon embodied in
the international trade of Brazil: an inputeoutput approach. Ecological
Economics 39, 409e424.
Miller, R., Blair, P., 1985. Inputeoutput Analysis: Foundations and Exten-
sions. Prentice Hall, New Jersey.
Ministry of Science and Technology Economy & Energy e NGO, 2001.
Supply of an Instrument for Estimating the Emissions of the Greenhouse
Effect Gases Coupled with the Energy Matrix. [is this a book, report?]
<http://ecen.com/matriz/eee24/coefycin.htm>.
Mu, H., Kondoua, Y., Tonookab, Y., Satoa, Y., Zhou, W., Ningb, Y.,
Sakamotod, K., 2004. Grey relative analysis and future prediction on rural
household biofuels consumption in China. Fuel Processing Technology 85,
1231e1248.
National Bureau of Statistics, P.R. China, 1998. China Statistical Yearbook
(1998). China Statistics Press, Beijing (in Chinese).
Ni, W.D., Thomas, B.J., 2004. Energy for sustainable development in China.
Energy Policy 32, 1225e1229.
Pan, H., 2005. The cost efficiency of Kyoto flexible mechanisms: a top-down
study with the GEM-E3 world model. Environmental Modelling & Soft-
ware 20, 1401e1411.
Paul, S., Bhattacharya, R.N., 2004. CO2 emissions from energy use in India:
a decomposition analysis. Energy Policy 32, 585e593.
Qu, G.P., 1992. China’s dual-thrust energy strategy for economic development
and environmental protection. Energy Policy 20, 500e506.
Roca, J., Alcantara, V., 2001. Energy intensity, CO2 emissions and the environ-
mental Kuznets curve. The Spanish case. Energy Policy 29, 553e556.
Sanchez-Choliz, J., Duarte, R., 2004. CO2 emissions embodied in international
trade: evidence for Spain. Energy Policy 32, 1999e2005.
Savabi, M.R., Stockle, C.O., 2001. Modeling the possible impact of increased
CO2 and temperature on soil water balance, crop yield and soil erosion.
Environmental Modeling & Software 16, 631e640.
Scrimgeour, F., Oxley, L., Fatai, K., 2005. Reducing carbon emissions?
The relative effectiveness of different types of environmental tax:
the case of New Zealand. Environmental Modelling & Software 20,
1439e1448.
Silberglitt, R., Hove, A., Shulman, P., 2003. Analysis of US energy scenarios:
meta-scenarios, pathways, and policy implications. Technological Fore-
casting & Social Change 70, 297e315.
Solveig, G., Wei, T.Y., 2005. Coal cleaning: a viable strategy for reduced
carbon emissions and improved environment in China? Energy Policy
33, 525e542.
Sun, J.W., 1999. The nature of the CO2 emissions Kuznets curve. Energy Pol-
icy 27, 691e694.
Sun, J.W., 2003. The natural and social properties of CO2 emissions intensity.
Energy Policy 31, 203e209.
Winiwarter, W., Schimak, G., 2005. Environmental software systems for emis-
sion inventories. Environmental Modelling & Software 20, 1469e1477.
Workgroup of the Academy of Macroeconomic Research of the State
Development Planning Commission, 1999a. Nuclear Power Station in
China’s Energy Structure. Study of a Macro Economy 8, 41e45.
(in Chinese).
Workgroup of the Academy of Macroeconomic Research of the State Devel-
opment Planning Commission, 1999b. Medium/Long-term Energy Strat-
egy for China. China Plan Publishing House, Beijing (in Chinese).
Wu, K., Li, B., 1995. Energy development in China. Energy Policy 23, 167e178.
Wu, L.B., Kaneko, S., Matsuoka, S., 2005. Driving forces behind the
stagnancy of China’s energy-related CO2 emissions from 1996 to 1999:
the relative importance of structural change, intensity change and scale
change. Energy Policy 33, 319e335.
Wu, Z.X., He, J.K., Zhang, A.L., Xu, Q., Zhang, S.Y., Sathaye, J., 1994.
A macro-assessment of technology options for CO2 mitigation in China’s
energy system. Energy Policy 22, 907e913.
Xu, Z.M., Cheng, G.D., Chen, D.J., Templet, P.H., 2002. Economic diversity,
development capacity and sustainable development of China. Ecological
Economics 40, 369e378.
Xue, X.M., 1998. Calculation and comparison study of CO2 emissions from in
China’s energy consumption. Environmental Protection 4, 27e28 (in
Chinese).
Yan, L.G., Kong, L., 1997. The present status and the future development of
renewable energy in China. Renewable Energy 10, 319e322.
Yabe, N., 2004. An analysis of CO2 emissions of Japanese industries during
the period between 1985 and 1995. Energy Policy 32, 595e610.
Yao, R.M., Li, B.Z., Steemers, K., 2005. Energy policy and standards for the
built environment in China. Renewable Energy 30, 1973e1988.
Zhang, Z.X., 2000. Decoupling China’s carbon emissions increase from
economic growth: an economic analysis and policy implications. World
Development 28, 739e752.
Zhou, F.Q., 2000. Discussion of Macro Assumptions in the Future Evolution
of the Chinese Energy System and the Impact of Energy Security Consid-
erations on Planning. Energy Research Institute, State Development
Planning Commission, Beijing, China.
Zhou, W.S., Yukio, Y., 1996. Numerical simulation of future energy demand
and technical option of CO2 emission mitigation in China. World
Environment 4, 38e41 (in Chinese).
393Y. Fan et al. / Environmental Modelling & Software 22 (2007) 378e393
Dr. Ying Fan is a professor at the Institute of Policy and Management, Chinese
Academy of Sciences, China. Her research field is energy and environment policy
andsystemengineering. In2004, shewasavisitingscholaratCornellUniversity,USA.
Ms. Qiao-Mei Liang is a Ph.D. candidate in Management Science at the
Institute of Policy and Management, Chinese Academy of Sciences, China.
Dr. Yi-Ming Wei is the deputy director and a professor at the Institute of
Policy and Management, Chinese Academy of Sciences, China. In 2005, he
was a visiting scholar at Harvard University, USA.
Dr. Norio Okada is a professor at the Urban Management Department of
Kyoto University, Japan.