Journal of Cleaner Production 103 (2015) 705e714

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Estimation of carbon dioxide emission in highway construction: a case

study in southwest region of China
Xianwei Wang a, Zhengyu Duan a, *, Lingsheng Wu a, b, Dongyuan Yang a
a
Key Laboratory of Road and Traffic Engineering of the Ministry of Education, School of Transportation Engineering, Tongji University, Shanghai 201804,
China
b
Sichuan Egang Highway Construction Co., Ltd, Chengdu, Sichuan 610041, China

a r t i c l e i n f o a b s t r a c t

Article history: The large-scale transportation infrastructure construction in developing countries such as China requires
Received 21 December 2013 emission estimation method for better project design. This study proposed an empirical method to es-
Received in revised form timate carbon dioxide (CO2) emission which was generated from highway construction based on four
7 October 2014
real projects in southwest region of China. The proposed method estimated the total emission from
Accepted 13 October 2014
Available online 22 October 2014
different steps of construction process (raw material production, material transportation and onsite
construction) by different project types (e.g. subgrade, pavement, bridge, and tunnels). The results show
that in general over 80 percent of the CO2 emission was generated from raw material production; the
Keywords:
Highway construction
onsite construction and material transportation only accounted for 10 and 3 percent of the whole CO2
Carbon dioxide emission emission, respectively. Moreover, the CO2 emission from bridge and tunnel constructions was much
CO2 emission larger than subgrade and pavement construction. The total CO2 emission from road, bridge and tunnel
Bridge construction constructions was 5229 kg/m, 35,547 kg/m and 42,302 kg/m, respectively. The empirical estimation
Tunnel construction method of the CO2 emission proposed in this study can be considered as references for CO2 emission
estimation in other regions which are similar as southwest region of China.
© 2014 Elsevier Ltd. All rights reserved.

1. Introduction Many studies have investigated the carbon emission issue in the
world. Dixit et al. (2010) identified the parameters for embodied
Global warming has been a severe challenge to the human being energy measurement for buildings, proposed the needs for an
in recent decades. It is widely known that the greenhouse gas, embodied energy measurement protocol for buildings (Dixit et al.,
especially carbon dioxide (CO2), blamed to be the main cause of 2012), and later proposed a conceptual model about system
global warming. According to U.S. Energy Information boundary for embodied energy (Dixit et al., 2013). Harmouche et al.
Administration (EIA), the total CO2 emission from the energy con- (2012) developed a carbon footprint calculator for building con-
sumption in the world was 32,578.645 million tons in 2011, among struction, with the consideration of project characteristics (e.g. size,
which 8715.307 million tons came from China (U.S. EIA, 2013). Such location, material choices). Han et al. (2013) also studied how to
amount of CO2 emission in China is believed to be generated by the embody energy consumption in building construction, and selected
large-scale transportation infrastructure construction. Since the a cluster of landmark commercial buildings in E-town, Beijing for
first expressway built in 1988 in China, the total length has been case.
96,200 km by the end of 2012, and the length will be increased to In the field of road engineering, some research frameworks and
108,000 km in 2015 (National Bureau of Statistics of China, 2013; emission estimation methods have been proposed and used for
Ministry of Transport of the People's Republic of China, 2011). environmental impact assessment. Park et al. (2003) assessed the
However, currently there are few empirical methods available environmental impacts of highways in life cycle. Fox et al. (2011)
which can be used to evaluate the CO2 emission estimation from introduced a pilot result of carbon management system for road
large-scale construction process in China. projects applied in Scotland. Tsai et al. (2012) proposed a frame-
work for developing sustainable items for highway design.
Avetisyan et al. (2012) proposed an optimization-based method-
* Corresponding author.
ology to permit a construction firm to assess its equipment needs
E-mail address: [email protected] (Z. Duan). while accounting for the greenhouse gas emissions. Huang et al.

http://dx.doi.org/10.1016/j.jclepro.2014.10.030
0959-6526/© 2014 Elsevier Ltd. All rights reserved.
706 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714

(2013) outlined the common methodology of road carbon foot pavement construction. Thanks to the large numbers of bridges and
printing, and performed case studies in UK and India using tunnels contained, general estimation results for these two struc-
CHANGER (the Calculator for Harmonized Assessment and tures were presented. In addition, the emission densities were
Normalization of Greenhouse-gas Emissions for Roads). And compared from the perspective of these three structures: road,
Barandica et al. (2013) evaluated Greenhouse gases in road con- bridge and tunnel. At last, we validated our results through the
struction from a life cycle perspective, and analyzed four con- comparison and discussion with similar researches and made final
struction projects in Spain. Specifically, the life cycle assessment of conclusion.
road pavement has been the focus by many researchers (Kim et al., The paper is organized as follows. Section 2 proposes the
2012; Cass and Mukherjee, 2011; Yu and Lu, 2012). Besides that, the empirical method in order to estimate CO2 emission. Section 3
University of California Pavement Research Center (UCPRC) has presents the case study of four real projects in southwest region
worked on recommending common practices for conducting LCA of China. Section 4 is the sensitivity analysis of the case study.
(Life Cycle Assessment) for pavements, and the UCPRC Pavement Section 5 is the results discussion, and Section 6 provides the
LCA Guideline intended for pavement LCA practitioners was pre- conclusions.
sented, which includes a LCA framework for pavements, as well as
some recommended data and models that have been used in Cal-
2. The proposed method for CO2 emission estimation
ifornia and other states in U.S. (Harvey et al., 2010). Michigan
Department of Transportation (DOT) also studied on developing
This study attempts to propose an empirical method to estimate
and implementing a project based life cycle framework which can
CO2 emission by using budget sheets, which record the detail
be used to estimate the carbon footprint for typical construction
consumptions of materials and energy. In the study scope, highway
work-items (Mukherjee and Cass, 2011). Moreover, Abolhasani
construction process is divided into three phases: raw material
et al. (2012) investigated the fuel use and emissions of non-road
production, material transportation and onsite construction. The
construction vehicles using real world data, and Frey et al. (2012)
total emissions are the sum of each stage:
summarized the characterization of real-world activity, fuel use,
and emissions of selected motor graders fueled with petroleum Q ¼ Q1 þ Q2 þ Q3 (1)
diesel and B20 biodiesel. Chang and Kendall (2011) made a life cycle
greenhouse gas assessment of infrastructure construction for Cal- Where, Q is the total amount of CO2 emission from highway con-
ifornia's high-speed rail system. struction; Q1, Q2, and Q3 is the amount of CO2 emission from raw
Although many studies attempted to estimate the carbon material production, material transportation and onsite construc-
emission in different regions, their results were not quite consis- tion, respectively. The unit of Q, Q1, Q2, and Q3 is kg.
tent, and there is a need for local researchers in China to propose its
own estimation method according to its own circumstances. Shang
et al. (2010) analyzed the energy consumption and atmospheric 2.1. The CO2 emission from raw material production
emissions of highway roadbed earthwork, drainage, protection and
pavement engineering in life cycle. Pan (2011) investigated the The CO2 emission from raw material production refers to the
energy consumption and carbon emissions of asphalt and cement CO2 emissions of materials which are generated in the mining,
concrete pavement in life cycle. Ouyang and Liu (2011) studied the processing and manufacturing factories and other production
carbon emissions of bridge project in life cycle with a case study. Xu processes. The formula to record CO2 emissions from raw material
(2012) calculated bridge cases' carbon emissions of two structures production can be defined as:
(e.g. concrete beams and steel - concrete composite beams). Li et al. Xn Xn
(2011) made a case study on the CO2 emissions from working Q1 ¼ i¼1
Q1i ¼ i¼1
ðqi  ui Þ (2)
machinery during highway tunnel construction.
Recently, the government of China started to realize to reduce Where, Q1 is the total CO2 emission from raw material production;
the large quantity of CO2 emission from the large-scale trans- Q1i is the CO2 emission from the production of material i; qi rep-
portation infrastructure construction. The first step of this work is resents the coefficient of CO2 emission from the production of
to quantitatively estimate CO2 emission by different types of material i; ui refers to the quantity of material i; i ¼ 1, 2, …, n, are the
transportation infrastructure construction (e.g. subgrade, pave- types of materials considered. The unit of Q1 and Q1i is kg.
ment, bridge, and tunnels). While lots of work have been done The major materials for highway construction are steel, cement,
worldwide in order to estimate of CO2 emission, however, there is asphalt and stone. The steel materialized process consists of two
still not a widely accepted method available. Moreover, the previ- stages: raw material exploitation and steel production. Direct and
ous studies usually focused on one type of road construction indirect environmental impacts may exist in either stage. There are
without considering subgrade, pavement, bridges and tunnels. several studies investigated the emission from the two stages. Gong
Currently there is still a need for the method of CO2 emission (2004) detailed analyzed the steel materialized system from
estimation that can be used in the real projects in China. Such exploitation to final product, in which direct environmental im-
method is potentially very helpful in the transportation infra- pacts from processing and transportation combined with the in-
structure construction. For example, a short-distance bridge/tunnel direct impacts from energy production were considered. Luo et al.
or long-distance road, which is more environmental friendly based (2011) also calculated the emission factor of steel production, but
on the evaluation of CO2 emission estimation? the value obtained was less than Gong (2004)'s, because Luo et al.
This study proposed an empirical method to estimate CO2 (2011) just considered the emission of material production pro-
emission from highway construction in China. First, the whole cessing part. He and Zhang (2013) built a model and discussed the
transportation infrastructure construction process was divided into carbon emissions and influencing factors of steel industry in China.
three steps: raw material production, material transportation and They concluded that the energy consumption and emissions had
onsite construction. Then the CO2 emission was calculated based on increased a lot in developing countries with the economy devel-
project budget sheet from four real highway projects in southwest opment. The extensive development mode is still common at cur-
region of China. After that the emissions from each phase and in- rent stage, but the energy-saving technology has been improved.
ternal compositions were detailed analyzed for subgrade and They found that the carbon emission density for steel production
 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714 707

Table 1 Table 3
Previous studies on CO2 emission factors of material production. The CO2 emission factors of electric power generation in China (2000e2009).

Material Author Scope Value Year 2000 2001 2002 2003 2004

Steel Gong (2004) Direct impacts from raw 4339 kg/t for large CO2 emission factor 0.834 0.797 0.805 0.816 0.801
material exploitation, shape steel; (kg/kw*h)
transportation and 3589 kg/t for light Year 2005 2006 2007 2008 2009
production process; shape steel; CO2 emission factor 0.792 0.783 0.766 0.723 0.723
indirect impacts from 3551 kg/t for linear (kg/kw*h)
the production of steel.
energy consumed
Wang (2007) Energy upstream phase 3493 kg/t
and material production techniques, as the factors of stone exploitation are negligible
process compared to steel and cement's, it will not materially influence the
Luo et al. Material production process 3545.6 kg/t final results. We rely on the CO2 emission factors from Zhang (2002)
(2011)
for stone materials. The CO2 emission factors of material production
Cement Gong (2004) Direct impacts from raw 1042 kg/t for
material exploitation, raw P.I. 52.5; used in this study are summarized in Table 2.
meal preparation, clinker, 920 kg/t for
grinding, packing and P.O. 42.5; 2.2. The CO2 emission from material transportation
transportation process; 678 kg/t for
indirect impacts from P.S. 32.5.
electric power production The material transportation means transferring materials from
and coal production related. factories to construction site using vehicles, which is an integral
Wang (2007) Energy upstream phase and About 1000 kg/t part of highway construction. CO2 emission from the transport part
material production process can be calculated as:
Xn Xn
Q2 ¼ i¼1
Q2i ¼ i¼1
ðui  di  ci Þ (3)
decreased by 10% from 2005 to 2009 by macro data, however, they
did not give the final emission factors.
Where, Q2 is the total amount of CO2 emission from raw material
The scope considered for CO2 emission from cement production
transportation; Q2i is the amount of CO2 emission from the trans-
in Gong (2004) includes the direct impacts from raw material
portation of material i; ui is the quantity of material i; di is the
exploitation, raw meal preparation, clinker, grinding, packing and
transport distance of material i; ci refers to the coefficient of CO2
transport process and indirect impacts from electric power pro-
emission for transferring material i. The unit of Q2 and Q2i is kg, and
duction and coal production related. The emission factor of cement
the unit of di is km.
production depends on the production skills, though skills of shaft
Emission factor ci is dependent much on the transportation
kiln and rotary kiln occupied the major position in about 2004,
method because of the difference in energy consumption efficiency
Gong (2004) selected the skill of new dry process for study. Xu
between different methods. Calculation models and related factors
(2013) studied the status of cement industry in China, indicated
can be found in Liu et al. (2010).
that the technique of new dry process is dominating in developed
countries, and suggested this process will soon replace the methods
2.3. The CO2 emission from onsite construction
of “shaft kiln” and “wet process kiln” in China.
Wang (2007) also analyzed the energy consumption and emis-
Emissions are produced by various types of mechanical equip-
sions of steel and cement production based on life cycle assess-
ment during onsite construction phase, which can be calculated
ment. Table 1 summarized the study scopes and results by different
using the following formula:
authors for CO2 emission factors during steel and cement
production. Xp Xp  
The results from Gong (2004) got the high priority for its com-
Q3 ¼ Q ¼
f ¼1 3f f ¼1
Af  Xf (4)
plete consideration scope, detailed analysis process and precise
calculation, and the results keep well consistent with others' Where, Q3 is the total CO2 emission from mechanical construction;
through comparison. CO2 emission factors of stone materials Q3f is the CO2 emission from machine f; Af represents the coefficient
exploitation were not mentioned in Gong (2004), but Zhang (2002) of CO2 emission of machine f; Xf is the operation time of machine f;
collected data and built the emission factor database according to f ¼ 1, 2 … P, are the various types of machines used for onsite
condition of Taiwan through his study of green building assess- construction. The unit of Q3 and Q3f is kg.
ment. Though CO2 emission from the production phase may vary in The factor Af can be estimated as:
different regions due to the variety of exploitation and processing
Af ¼ Ef  Mf (5)

Table 2 Where, Ef represents the coefficient of CO2 emission from the en-
The CO2 emission factors of main material production.
ergy combusted by machine f; Mf is the quantity of energy
Material type CO2 emission factor Unit consumed per working time for machine f.
Large shape steel 4339 kg/t
Light shape steel 3589 kg/t
Linear steel 3551 kg/t Table 4
Cement (P.I.52.5) 1042 kg/t The CO2 emission factors of main energy sources.
Cement (P.O.42.5) 920 kg/t
Energy CO2 Emission factor Unit
Cement (P.S.32.5) 678 kg/t
Asphalt 248 kg/t Fuel oil 3.241 kg/kg
Emulsified asphalt 160 kg/t Gasoline 2.988 kg/kg
Pebbles 3.11 kg/m3 Diesel 3.164 kg/kg
Gravel 3.9 kg/m3 Electricity 0.723 kg/kw*h
708 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714

from 2000 to 2009 (transmission losses included), and the results

are presented in Table 3.
It can be seen that CO2 emission factor of electric power
decreased from 2000 to 2009, except that the factors of 2002, 2003,
and 2004 were more than that of 2001, and it remained the same in
2008 and 2009. According to the trend of changing, ‘0.723’ was
adopted in this case study, and further discussion will be provided
in sensitivity analysis. So the CO2 emission factors of main energy
sources are presented in Table 4 below.

2.4. Data collection and calculation flowchart

Data was collected from project budget sheets in this study,

which were obtained from highway construction company, with
detailed and complete records of highway construction process.
Each sheet includes detailed records of each type of constructions
(subgrade, pavement, bridges and tunnels), quantity and price of
raw materials, workdays and energy consumptions of machines.
The calculation process is illustrated in Fig.1.

Fig. 1. Flowchart of CO2 emission estimation of highway construction.

3. Case study

Machinery in highway construction mainly include vehicles, 3.1. Background of the highway projects
bulldozers, mobile air compressors, rollers, asphalt mixing
equipment, loaders, AC arc welder, electric winches, cranes, According to the national strategy to promote the economy
electric air compressors, ventilators, etc. And the main energies development in the central and western regions in China, many
consumed include fuel oil, gasoline, diesel and electricity. transportation infrastructures have been planned to be con-
Luo et al. (2011) took advantage of the emission estimation structed. The west region of China is largely covered by hill and
method from IPCC (2006), and calculated the emission factors from mountains, making the highway routing and design a more com-
the combustion of fuel oil, gasoline and diesel based on the net plex problem. Furthermore, in the stage of design, more efforts
heating value of each item according to the condition of China. need to be paid for the consideration of the ecological protection
Since electricity may be generated by thermal power, hydropower, and reduction of carbon emissions.
nuclear power, wind, geothermal, tidal and solar power, and In this case study, three newly built highway projects were
emission mainly come from thermal power, the emission factor selected, located in southwest region of China, from Suining to
from electricity generation is closely related to the energy struc- Guang'an, Xuyong to Gulin and Zigong to Longchang respectively
ture. Luo et al. (2011) calculated the CO2 emission factor from (Fig. 2.). The three highways share a similar structural design, with
electric power generation based on the energy structure in China a designed speed of 80 km/h, four lanes, subgrade width of 24.5 m

Fig. 2. Locations of the projects in case study (original picture obtained from Google Map).
 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714 709

Table 5 Table 7
Characteristics of selected projects and the results from questionnaire survey. The CO2 emission estimation of subgrade construction.

Projects Suinig to Xuyong Zigong to General results from Project 1 Project 2 Project 3 Project 4
Guang'an to Gulin Longchang questionnaire
Length (km) 68.84 6.62 17.40 40.06
Terrain Heavy hill Mountain Heavy hill All types of terrain Total emissions (t) 110,356 47,933 85,274 32,305
been covered Material production (t) 55,692 37,502 56,090 9632
Proportion 14.69% 56.92% 9.57% 63% below 30%; 31% (Percentage among total) (50.5%) (78.2%) (65.8%) (29.8%)
of bridge between 30% and Material transport (t) 5004 725 1230 No data
and tunnels 50%; 6% between (Percentage among total) (4.5%) (1.5%) (1.4%)
50% and 70% Onsite construction (t) 49,659 9706 27,954 22,672
Non-linear 1.14 1.25 1.12 37% between 1.0 and (Percentage among total) (45.0%) (20.2%) (32.8%) (70.2%)
coefficient 1.2; 31% between 1.2 Emission density (t/km) 1603 7238 4901 806
and 1.4; 16% between
1.4 and 1.6; 16%
above 1.6
Proportion 20.60% 39.29% 19.21% 33% between 15% and simplify the analyses, in the following paper, the project from
of special 25%; 47% between Suining to Guang'an is named as Project 1, the first part of Xuyong
subgrade 25% and 35%; 20% to Gulin as Project 2, the second part of Xuyong to Gulin as Project 3
between 35% and 45%
and the project of Zigong to Longchang as Project 4.
Transport 80 km 100 km / 7% below 50 km; 33%
distance between 50 km and
of materials 100 km; 40% 3.2. Total CO2 emission estimation of highway construction
between 100 km and
150 km; 20% above
Table 6 presents the total CO2 emission estimation of the four
150 km
Thickness of 56 cm 56 cm 56 cm 6% below 40 cm;
selected projects by using the proposed methodology. The results
base and 37% between 40 cm show that major part of emissions, over 80 percent, came from raw
subcase level and 55 cm; 44% material production. The emissions from material transportation
between 55 cm and account for the least part, only less than 3 percent. On average,
70 cm; 13%
about 12 percent of emissions came from onsite construction.
between 70 cm and
85 cm The emission densities (emissions per kilometer) ranged from
Thickness 18 cm 18 cm 18 cm 44% between 10 cm 5609 t/km to 29,677 t/km in the four selected projects, and can be
of asphalt and 15 cm; 56% grouped into two categories: densities of projects 2 and 3 are
surface between 15 cm and
significantly higher than those of projects 1 and 4. Furthermore, the
20 cm
emission densities seemed to be closely associated to the pro-
portions of bridges and tunnels in the project, therefore emission
estimations of roads, bridges and tunnels were discussed respec-
and asphalt concrete pavement. In order to verify the representa- tively in the following sections.
tives of the chosen cases, a questionnaire survey was conducted
among more than 30 engineers, consulting about the characteris- 3.3. The CO2 emission estimation of subgrade and pavement
tics of highway project in general, including the terrain, proportion constructions
of bridge and tunnel, non-linear coefficient, distance of material
transport, etc. Table 5 shows the characteristics of the projects and This section analyzes the estimation of CO2 emission of sub-
the general results from the questionnaire survey. grade and pavement constructions. Subgrade is the foundation of
Please note that the project of Xuyong to Gulin was divided into highway and the associate cut and fill works require lots of ma-
two parts in the design process. So in this study, we estimated the chines as excavators, bulldozers, loading and unloading cars, rollers
emission separately for the two parts of the project. In order to and air compressors. Pavement construction requires materials like
sand, gravel, cement, asphalt and steel, and machines of asphalt

Table 6
The estimation of CO2 emission of the four selected projects.

Project 1 Project 2 Project 3 Project 4

Total length (km) 97.455 23.698 41.140 50.578

Design speed (km/h) 80
Number of lanes 4
Subgrade width (m) 24.5
Pavement structure Asphalt concrete pavement
Number of bridges 73 35 54 25
Total length of bridges (km) 14.315 8.438 15.880 4.840
Number of tunnels None 7 6 None
Total length of tunnels (km) None 7.408 4.710 None
Total emissions (t) 741,033 703,287 1,091,184 283,691
Material production (t) 626,187 627,267 982,091 240,557
(Percentage among total) (84.50%) (89.20%) (90.00%) (84.80%)
Material transportation (t) 17,624 6830 11,989 No
(Percentage among total) (2.40%) (1.00%) (1.10%) data for
transportation
Onsite construction (t) 97,222 69,190 97,104 43,134
(Percentage among total) (13.1%) (9.8%) (8.9%) (15.2%)
Emission density (t/km) 7604 29,677 26,524 5609
Proportion of bridges and 14.69% 66.87% 50.05% 9.57%
Fig. 3. The CO2 emission estimation of different types of work in subgrade
tunnels in the project
construction.
710 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714

Table 8 Table 10
The CO2 emission estimation of pavement construction. The CO2 emission from pavement construction of road, bridge and tunnel.

Project 1 Project 2 Project 3 Project 4 Project 1 Project 2 Project 3 Project 4

Length(km) 83.151 22.468 37.990 44.900 Road pavement (km) 68.836 6.622 17.400 40.060
Total emissions (t) 125,154 13,918 33,904 52,034 Bridge pavement (km) 14.315 8.438 15.880 4.840
Material production (t) 92,829 9477 2359 41,388 Tunnel pavement (km) 0 7.408 4.71 0
(Percentage among total) (74.2%) (68.1%) (71.8%) (79.5%) Emission of road 115,071 10,407 27,876 60,929
Material transportation (t) 7,014 704 1,809 No data pavement (t)
(Percentage among total) (5.6%) (5.1%) (5.3%) Emission of bridge 3119 1624 3331 1008
Onsite construction (t) 25,311 3,737 7,735 10,646 pavement (t)
(Percentage among total) (20.2%) (26.9%) (22.8%) (20.5%) Emission of tunnel 0 1190 898 0
Emission density (t/km) 1505 619 892 1159 pavement (t)
Emission density of road 1672 1572 1602 1521
pavement (t/km)
Emission density of bridge 218 192 210 208
mixing equipment, dump trucks, loaders and road rollers. Emis- pavement (t/km)
sions are generated from all these factors. Table 7 shows the Emission density of tunnel None 161 191 None
emission estimation of subgrade construction and its composition. pavement (t/km)
It can be observed that the emissions of raw material production
and onsite construction accounted for more than 95 percent.
The emissions composition from different types of work for while the widths of all bridges are around 25 m. According to the
subgrade construction is presented in Fig. 3. Cut and fill work, overview emission estimation, material production and onsite
special subgrade treatment, and protection work were found to be construction were the main sources of CO2 emissions, coupled with
the main sources of the emissions. Combining Table 7 and Fig. 3, it the reason that the transportation data were missed in Project 4, so
can be inferred that special subgrade treatment of Project 2 and 3 the emissions of transportation were not considered in this section.
consumed lots of materials, significantly raising the emissions in As there are 187 bridges considered here, we try to analyze the
materials production and making emission densities higher than changing trend of the emissions and give a more universal
other projects. So the emissions may increase greatly for highway conclusion instead of analyzing emission composition of one
constructed in adverse terrain. project only. As the emission factor for gravel is too little, the
Table 8 shows the CO2 emission estimation of pavement con- emissions can be neglected compared to steel and cement. Only
struction. It can be observed that the raw material production and steel, cement and onsite construction are taken into consideration
onsite construction are the two major sources of emission, here for bridges' emissions. The results were depicted in Fig. 5,
furthermore, the emissions of raw material production itself which shows with the increase of the bridges' length, the emissions
accounted for more than 70 percent among total. of materials seem grow more rapidly than those of onsite con-
Fig. 4 shows the emission estimation by different types of work struction, and the emissions of steel increase more rapidly than
in pavement construction. The road sub-base, base and asphalt those of cement. So steel production is the most contributor of
concrete pavement generate the most emissions, probably because emissions, cement is the second source and onsite construction
of a large quantity of materials used. ranks the third. All the emissions' growth rates of these three cat-
Based on Fig. 4, each project seems to share a similar compo- egories (steel, cement, onsite construction) reached a mutation
sition of emissions from different types of works, while the emis- when the bridges' length exceeds 600 m, i.e. the growing rates
sion densities are totally different (Table 8). This might be that became much faster. It may be inferred that the quantity of mate-
projects of road, bridge and tunnel were put together when pave- rials consumed and the machine working hours required for
ment is considered, and it will be explained further in the following bridges per meter increase in a more sharp way when the length
comparison section (Table 10). reaches a certain amount, probably due to different structures must
be applied for bridges overpass a certain length.
3.4. The CO2 emission estimation of bridge construction A good understanding of possible emission density value is very
useful to project designers. Hence the relationship between CO2
3.4.1. Total CO2 emission estimation of bridge construction emission densities with lengths of bridges is presented in Fig. 6. We
In this study, 187 bridges were selected in total, among which noticed that the dots gathered at the range from 20,000 kg/m to
the shortest is about 20 m long, and the longest is about 1000 m, 40000 kg/m when the bridges were at small scale, under the length

Table 9
The estimation of CO2 emission of tunnels.

Tunnels Length CO2 emission of Proportion CO2 emission of steel Proportion CO2 emission of Proportion Total CO2
(m) onsite construction (kg) production (kg) cement production (kg) emission (kg)

Separated tunnel 1 409 1,749,313 14.3% 4,024,793 33.0% 6,417,948 52.6% 12,192,054
Separated tunnel 2 492 2,371,619 13.6% 6,442,515 36.9% 8,630,262 49.5% 17,444,397
Separated tunnel 3 3950 19,231,063 15.2% 32,755,208 25.9% 74,289,138 58.8% 126,275,409
Separated tunnel 4 950 4,506,904 14.2% 9,180,989 28.9% 18,067,344 56.9% 31,755,238
Separated tunnel 5 321.5 1,610,970 13.5% 4,046,865 33.9% 6,289,128 52.6% 11,946,963
Separated tunnel 6 1112.5 5,457,608 13.9% 12,976,679 33.1% 20,763,750 53.0% 39,198,037
Separated tunnel 7 2760 14,276,807 14.2% 37,283,008 37.1% 48,801,084 48.6% 100,360,899
Double arch tunnel 1 345 2,065,032 11.8% 8,014,069 45.8% 7,424,778 42.4% 17,503,880
Double arch tunnel 2 320 1,707,685 11.1% 6,355,982 41.5% 7,257,312 47.4% 15,320,979
Double arch tunnel 3 239 1,373,328 11.0% 5,480,753 43.8% 5,657,232 45.2% 12,511,313
Double arch tunnel 4 387 2,297,692 11.6% 8,531,682 43.1% 8,948,244 45.2% 19,777,619
Double arch tunnel 5 396 2,319,764 11.4% 8,789,933 43.1% 9,276,396 45.5% 20,386,094
Double arch tunnel 6 434 2,534,592 10.9% 10,471,470 45.2% 10,153,050 43.8% 23,159,113
 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714 711

Fig. 4. The CO2 emission estimation of different types of work in pavement Fig. 6. The estimation of CO2 emission density of bridge construction.
construction.

concrete T beam bridge were adopted for bridges 100 me600 m

of 600 m, though there were a few cases whose emission densities in length, and the emission density varied from about 30,000 kg/m
were far more than that. However, when the bridge length excee- to more than 40,000 kg/m, but might reach 70,000 kg/m in some
ded 600 m, the values of CO2 emission densities all became larger, special cases. The pre-stressed concrete continuous rigid fame type
even surpassed 100,000 kg/m. It might be inferred that the emis- suits large scale bridges, and the emission densities stayed above
sion density much dependent on different bridge structures, so 60,000 kg/m, with some even surpassed 100,000 kg/m.
further discussion is necessary.

3.4.2. The CO2 emission density of bridges by different structure 3.5. The CO2 emission estimation of tunnel construction
types
Bridges can be generally classified as beam bridge, arch bridge, Tunnel construction needs a large amount of steel and cement.
suspension bridge, and cable-stayed bridges by structure type. The onsite machinery includes electric air compressor, ventilators,
Beam bridges are the most common type used today. All the dump trucks and mixers, etc. The tunnels in this case study can be
selected bridges belong to beam bridge type, but they can be classified as separated tunnels and double arch tunnels by structure
further classified as reinforced concrete hollow board, pre-stressed type. Thirteen tunnels are included, among which there are seven
concrete box girder, pre-stressed concrete T beam and pre-stressed separated tunnels and six double arch ones. The length of each
concrete continuous rigid frame. The carbon dioxide emission tunnel, total CO2 emissions, emissions from steel, cement, onsite
densities of these four types of bridges were shown in Fig. 7. We can construction and their proportions are presented in Table 9.
notice that each structure type got a suitable bridge scale, and the The proportions of CO2 emission from onsite construction,
emission densities differed obviously among them. The reinforced production of steel and cement keep comparatively steady at
concrete hollow board type seems to be more suitable for short different tunnel scales, but there is an obvious difference between
bridges, and the emission densities range from about 30,000 kg/m separate tunnels and arch ones. For separated tunnels, the CO2
to about 40,000 kg/m. The bridge scales of pre-stressed concrete emission from onsite construction accounted for about 14%, emis-
box girder type range from about 30 m to about 400 m, and its sion of steel production contributed between 25% and 37%, and
emission densities were around 30,000 kg/m. The pre-stressed emission of cement comprised around 50%. However, for double
arch tunnels, onsite construction emission accounted for about 11%,

Fig. 5. The composition of CO2 emission of bridge construction: steel production, Fig. 7. The estimation of CO2 emission density of bridge construction by different
cement production and onsite construction. structure types.
712 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714

5229 kg/m, for bridge is 35,547 kg/m, and for tunnel is 42,302 kg/m
(Fig. 9).

4. Sensitivity analysis

Sensitivity analysis is necessary to assess the impact of uncer-

tain factors on CO2 emission. According to the analysis above, there
might be a timeliness problem for CO2 emission factors of steel,
cement and electric power production, steel and cement are both
the most important sources of emissions in highway construction.
Therefore, CO2 emission factors of steel production, cement pro-
duction and electricity generation are set as control variables, and
influence on different stages and on emission density of different
structures are analyzed thereafter.
Fig. 8. The estimation of CO2 emission density of tunnel constructions. With the expectation for the energy-saving techniques applied
in steel and cement production, we assume to reduce them by 10%.
According to the study by Luo et al. (2011), we expect the electricity
emission of steel contributed for more than 40%, and emission of production factor to decrease by 5%.
cement comprised about 45%. Table 11 shows that the changes happened on different emission
In order to compare the CO2 emission levels of these tunnels stages caused by the adjustment of factors were rather small. When
more clearly, emission densities were mapped out in Fig. 8. The the emission factor of steel production was decreased by 10%, the
densities of each type of tunnel appear to be relatively stable, while emission proportions of material production dropped 0.6%, those of
the difference between the two types is rather obvious. More onsite construction increased 0.4%e0.6%, and those of transport
specifically, the CO2 emission densities of separated tunnels valued raised no more than 0.1%. When emission factor of cement pro-
from about 30,000 kg/m to about 37,000 kg/m, while those of duction was decreased by 10%, the emission proportion of materials
double arch tunnels were around 50,000 kg/m. production dropped about 0.4%e0.7% and onsite construction
increased 0.3%e0.6%. The impacts of electricity factor's adjustment
3.6. Comparison of CO2 emission density between road, bridge and were even more negligible, only 0.1%e0.2% difference in both ma-
tunnel constructions terial production and onsite construction.
Table 12 show the results of sensitivity analysis of CO2 emission
In this section, we attempt to compare the CO2 emission density densities of different structures. The results show that when the
of road, bridge and tunnel construction. Firstly, road, bridge and emission factor of steel production decreased by 10%, the CO2
tunnel need to be separately analyzed as for the CO2 emission emission density of road dropped 2.91%, the emission density of
estimation of pavement. The values of emission density were quite bridge fell 5.53%, and that of tunnel declined by 3.40%. The changing
close in the same category but differed enormously between cat- proportion to each structure seems to be associated with the
egories (Table 10). quantity of material used in the structure. When the emission factor
Then, the CO2 emissions of subgrade and road pavement were of cement decreased by 10%, the emission density of road dropped
included in road, the emissions of bridge engineering and bridge 3.40%, the emission density of bridge fell 2.77%, and that of tunnel
pavement were included in bridge, and the emissions of tunnel declined 5.03%. On the other hand, little changes were caused by the
engineering and tunnel pavement were included in tunnel. Ac- adjustment of electricity factor. When the emission factor of elec-
cording to the above estimation, the average emission value is tricity generation was decreased by 5%, the emission density of road,
3637 kg/m for subgrade engineering, 1592 kg/m for road pavement, bridge and tunnel fell 0.13%, 0.18% and 0.48%, separately.
207 kg/m for bridge pavement, 176 kg/m for tunnel pavement,
35,340 kg/m for bridge engineering, and 42,126 kg/m for tunnel Table 11
engineering. Consequently, the average CO2 emission for road is Sensitivity analysis of the estimation of CO2 emission on different stages.

Factor adjustment Impacts on Project 1 Project 2 Project 3 Project 4

stages

Emission factor Material 0.63% 0.60% 0.52% 0.64%

of steel production
production (10%) Material 0.10% 0.00% 0.06% Null
transportation
Onsite 0.53% 0.60% 0.46% 0.64%
construction
Emission factor Material 0.67% 0.59% 0.40% 0.62%
of cement production
production (10%) Material 0.10% 0.00% 0.04% Null
transportation
Onsite 0.56% 0.59% 0.36% 0.62%
construction
Emission factor Material 0.12% 0.16% 0.21% 0.16%
of electricity production
generation (5%) Material 0.00% 0.04% 0.00% Null
transportation
Onsite 0.13% 0.12% 0.21% 0.16%
construction

Fig. 9. The estimation of CO2 emission density of road, bridge and tunnel Note: the percentage values in the table are the direct difference between the results
constructions. after adjustment with the initial values.
 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714 713

Table 12 endogenous CO2 emission from the tunnel construction (just

Sensitivity analysis of CO2 emission densities of different structures. considering the emissions of fossil fuel combustion of construction
Factor adjustment Road Bridge Tunnel machinery), and the result was about 3000 kg/m. According to the
Emission factor of steel production (10%) 2.91% 5.53% 3.40%
proportion of onsite construction part may account for among total
Emission factor of cement production (10%) 3.40% 2.77% 5.03% emissions as we estimated, the results keep a good consistency.
Emission factor of electricity generation (5%) 0.13% 0.18% 0.48%

6. Conclusion
5. Result discussion
This study proposed an empirical method to estimate the Car-
Due to the difference of the study scope, the comparison of CO2 bon Dioxide (CO2) emission based on four selected projects of
emission estimation should be very carefully (Barandica, 2013).In highway construction in southwest region of China. We divided the
this study, the result shows that over 80 percent of the total construction period into three stages: material production, material
emission came from raw material production for the selected transportation and onsite construction. The results show that the
projects, while only 10 percent and 3 percent of the total emissions major CO2 emission was generated from the material production,
came from the onsite construction and material transportation.. secondly from onsite construction, and transportation stage
This result is consistent with the analyses in previous studies (Park accounted for the least.
et al., 2003; Barandica, 2013; Cass and Mukherjee, 2011; Loijos, In this study, based on the collected data from 187 bridges and
2011). This indicates that the focus should be strongly put on ma- 13 tunnels, this paper proposed the general CO2 emission densities
terials production process when low-carbon is considered. for construction of three structure types (e.g. road, bridge and
Advanced techniques developed to decrease the emissions in ma- tunnel). The estimated emission density of road, bridge and tunnel
terial production have the most potential. is 5229 kg/m, 35,547 kg/m and 42,302 kg/m, respectively.
In this study, the results also show that the cut & fill work and The results in this study may benefit the highway design in
protection & reinforcement work generated large proportion of the China, especially in the southwest region. In the future, a supple-
emissions in subgrade construction. In addition, the special sub- mentary index may be added for overall assessment of projects. In
grade treatment has significant impact on the composition of addition, the detailed guidance for construction operations (e.g.
emissions. For pavement construction, the material production, equipment selection, green construction techniques) may also be
onsite construction and material transportation generated about considered in future.
70%, 20%, and 5% of the total emission, respectively. The emission
results from pavement construction were validated by the consis-
References
tency with the results obtained in Pan (2011). Barandica (2013)
found pavement generated 1.5%e5% proportions among the total Abolhasani, S., Frey, H.C., Kim, K., Rasdorf, W., Lewis, P., Pang, S., 2012. Real-world
emission; similarly in this study, the emission density from pave- in-use activity, fuel use, and emissions for nonroad construction vehicles: a case
ment appears rather less compared to subgrade, and more obvious study for excavators. J. Air Waste Manag. Assoc. 58 (8), 1033e1046.
Avetisyan, H.G., Miller-Hooks, E., Melanta, S., 2012. Decision models to support
compared to bridges or tunnels. The significant importance of greenhouse gas emissions reduction from transportation construction projects.
emission from earthwork found in this study is also highlighted by J. Constr. Eng. Manag. 138 (5), 631e641.
Barandica, J.M., Ferna ndez-Sa  Delgado, J.A., Acosta, F.J., 2013.
nchez, G., Berzosa, A.,
Barandica (2013). The reason inferred by Barandica (2013) was the
Applying life-cycle thinking to reduce greenhouse gas emissions from road
fact that Spanish orography was comparatively complex and a high projects. J. Clean. Prod. 57, 79e91.
number of embankments were needed, which is consistent with Cass, D., Mukherjee, A., 2011. Calculation of greenhouse gas emissions for highway
the terrain in this study. construction operations by using a hybrid life-cycle assessment approach: case
study for pavement operations. J. Constr. Eng. Manag. 137 (11), 1015e1025.
The comparison for emissions from bridges and tunnels is very
Chang, B., Kendall, A., 2011. Life cycle greenhouse gas assessment of infrastructure
difficult, since as far as we know, there are few studies about the construction for California's high-speed rail system. Transp. Res. Part D: Transp.
emission estimation of bridges and tunnels. Environ. 16 (6), 429e434.
Dixit, M.K., Culp, C.H., Fernandez-Solis, J.L., 2013. System boundary for embodied
The emissions of bridges construction were correlated with the
energy in buildings: a conceptual model for definition. Renew. Sustain. Energy
total length of bridges. Moreover, the CO2 emission densities Rev. 21, 153e164.
gathered at the range from 20,000 kg/m to 40000 kg/m when the Dixit, M.K., Fernandez-Solis, J.L., Lavy, S., Culp, C.H., 2010. Identification of param-
bridges were at small scale, under the length of 600 m; when the eters for embodied energy measurement: a literature review. Energy Build. 42,
1238e1247.
bridge length exceeded 600 m, the values of CO2 emission den- Dixit, M.K., Fernandez-Solis, J.L., Lavy, S., Culp, C.H., 2012. Need for an embodied
sities all became larger, even surpassed 100,000 kg/m. Further- energy measurement protocol for buildings: a review paper. Renew. Sustain.
more, we analyzed the bridges' emission density for different Energy Rev. 16, 3730e3743.
Fox, J., Bell, D., Edmond, G., Cummings, P., Langstraat, J., 2011. A practical tool for
structure types. low-carbon road design and construction. Transport 164, 165e179.
Xu (2012) found that the emission densities of bridges con- Frey, H.C., Kim, K., Pang, S., Rasdorf, W.J., Lewis, P., 2012. Characterization of real-
struction were 82,560 kg/m for material production, and 5740 kg/ world activity, fuel use, and emissions for selected motor graders fueled with
petroleum diesel and B20 biodiesel. J. Air Waste Manag. Assoc. 58 (10),
m for onsite construction, the sum of which was 88,300 kg/m. We 1274e1287.
got the value of 35,547 kg/m in this study. It is note that the width Gong, Z., 2004. A Quantitative Method to the Assessment of the Life Cycle. Master
of bridges in Xu (2012) is around 50 m, whereas the width of thesis. Tsinghua University, Beijing, China.
Han, M.Y., Chen, G.Q., Shao, L., Li, J.S., Alsaedi, A., Ahmad, B., Guo, S., Jiang, M.M.,
bridge is about 25 m in this study, so the results would be rather
Ji, X., 2013. Embodied energy consumption of building construction engineer-
comparable with this considered. Furthermore, due to the large ing: case study in E-town Beijing. Energy Build. 64, 62e72.
number of bridges in this study, a more reliable result is likely to Harmouche, N., Ammouri, A., Srour, I., Chehab, G., Hamade, R., 2012. Developing a
carbon footprint calculator for construction buildings. Constr. Res. Congr. 2012,
be obtained.
1689e1699.
In the tunnel cases estimated, the CO2 emission composition Harvey, J., Kendall, A., Lee, I.S., Santero, N., Van Dam, T., Wang, T., 2010. Pavement
proportions keep steady at different scales, but differed between Life Cycle Assessment Workshop: Discussion Summary and Guidelines. Tech-
separate tunnels and arch ones. The emission density of the sepa- nical memorandum: UCPRC-TM-2010-03. University of California, Davis, Davis,
California.
rate tunnels was between 30,000 kg/m and 37,000 kg/m, while He, W., Zhang, K., 2013. The decomposition analysis on the influencing factors of
50,000 kg/m for arch tunnels. Li et al. (2011) estimated the China's steel industry carbon emission. J. Industrial Technol. Econ. 1, 002.
714 X. Wang et al. / Journal of Cleaner Production 103 (2015) 705e714

Huang, Y., Hakim, B., Zammataro, S., 2013. Measuring the carbon footprint of road Ouyang, D., Liu, M., 2011. Calculation method of life cycle carbon emissions on
construction using CHANGER. Int. J. Pavement Eng. 14 (6), 590e600. bridges. J. Civ. Archit. Environ. Eng. 33, 125e129.
IPCC, 2006. Prepared by the national greenhouse gas inventories programme. In: Park, K., Hwang, Y., Seo, S., Seo, H., 2003. Quantitative assessment of environmental
Eggleston, H.S., Buendia, L., Miwa, K., Ngara, T., y Tanabe, K. (Eds.), 2006 IPCC impacts on life cycle of highways. J. Constr. Eng. Manag. 129 (1), 25e31.
Guidelines for National Greenhouse Gas Inventories. Institute for Global Envi- Pan, M., 2011. The Methodology Research and Application on Energy Consumption
ronmental Strategies (IGES), Japan. and Carbon Emissions of Highway Based on the Life Cycle Assessment. Master
Kim, B., Lee, H., Park, H., Kim, H., 2012. Framework for estimating greenhouse gas thesis. South China University of Technology, Guangzhou, China.
emissions due to asphalt pavement construction. J. Constr. Eng. Manag. 138 (11), Shang, C., Zhang, Z., Li, X., 2010. Research on energy consumption and emission of
1312e1321. life cycle of expressway. J. Highw. Transp. Res. Dev. 27, 149e154.
Liu, M., Qian, F., Zhan, X., 2010. Calculation model for energy carbon emission of Tsai, C.Y., Chang, A.S., 2012. Framework for developing construction sustainability
building material transportation. E-Product E-service and E-entertainment items: the example of highway design. J. Clean. Prod. 20, 127e136.
(ICEEE), 2010 International Conference on. IEEE, pp. 1e3. U.S. Energy Information Administration. International Energy Statistics: Total Car-
Li, X., Liu, J., Xu, H., Zhong, P., Zheng, X., Wang, Z., 2011. Calculation of endogenous bon Dioxide Emissions from the Consumption of Energy. Available from: http://
carbon dioxide emission during highway tunnel construction: a case study. In: www.eia.gov/cfapps/ipdbproject/IEDIndex3.cfm?tid¼90&pid¼44&aid¼8
Water Resource and Environmental Protection (ISWREP), 2011 International (accessed 13.07.05.).
Symposium on, vol. 3. IEEE, pp. 2260e2264. Wang, J., Zhang, X., Huang, Z., 2007. Life cycle assessment energy consumption and
Loijos, A., 2011. Life Cycle Assessment of Concrete Pavements: Impacts and Op- pollutant emission inventory analysis of construction materials production. Res.
portunities. Doctoral dissertation, Massachusetts Institute of Technology. Environ. Sci. 20 (6), 149e153.
Luo, Z., Yang, L., Han, B., 2011. Research on CO2 emission calculation method and Xu, S., 2012. The Research on Carbon Emissions of Different Structural Material in
CO2 reduction strategies of building materials. Build. Sci. 27, 1e8. Bridge Life Cycle. Master thesis. Wuhan University of Technology, Wuhan,
Ministry of Transport of the People's Republic of China, 2011. The Twelfth Five-year China.
Development Plan for Transport. Ministry of Transport of the People’s Republic Xu, X., 2013. Life Cycle Assessment of Cement in China. Master thesis. Dalian Uni-
of China, Beijing, China. versity of Technology, Dalian, China.
Mukherjee, A., Cass, D., 2011. Carbon Footprint for HMA and PCC Pavement. RC- Yu, B., Lu, Q., 2012. Life cycle assessment of pavement: methodology and case Study.
1553. Michigan Tech Transportation Institute. Transp. Res. Part D 17 (5), 380e388.
National Bureau of Statistics of China. Available from: http://data.stats.gov.cn/ Zhang, Y., 2002. Evaluation of Carbon Dioxide Reduction of Life Cycle for Buildings.
workspace/index;jsessionid¼8B4C40165D08CFA896A0D924430D9382? PhD thesis. National Cheng Kung University, Taiwan.
m¼hgnd (accessed 14.03.16.).