LCCA and environmental LCA for highway pavement selection in Colorado

Abstract

The roadway network in the USA earned a grade of D representing poor condition in the latest report card from the American Society of Civil Engineers. To maintain economic and environmental sustainability during the roadway network development and rehabilitation, it is critical to apply sustainable materials and intelligent design. A good estimation on project-level life-cycle costs and environmental impacts is one of the important steps in the highway investment decision-making process. This article examines the current life-cycle cost analysis (LCCA) practice employed by the Colorado Department of Transportation (CDOT) in their pavement investment decision-making process, and proposes a regional environmental life-cycle assessment (LCA) model to evaluate the greenhouse gas (GHG) emissions associated with Colorado highway pavements. Both LCCA and LCA are performed for a highway reconstruction project with Portland cement concrete pavement (PCCP) and hot-mixed asphalt (HMA) alternatives. The LCCA is 7.4% in favour of HMA. Since the difference is less than 10%, it indicates equivalent designs. However, in the LCA, the GHG emission from PCCP is 26% less than the HMA over the 40-year analysis period. The vehicle fuel consumption will increase due to the deterioration of pavements. But the increased user cost is not included in the current LCCA employed by CDOT as well as user cost due to crashes and nonuser costs. The LCA can be an optional criterion for the selection of the preliminary pavement type.

Keywords::

• pavement
• LCCA
• LCA

1. Introduction

There are approximately 4.2 million kilometres (2.6 million miles) of paved roads in the USA, which cope with an approximate volume of traffic of 13 billion vehicle-kilometres (8.2 billion vehicle-miles) per day (FHWA 2008; USDOT 2008). The US roadway network earned a grade of D representing a poor condition in the latest 2013 ASCE report card. Substantial investment is required every year in this network for maintenance and new construction. In addition, significant energy and resources are consumed annually, and millions of tons of greenhouse gases (GHGs) are released during the construction and operation of pavements. The US Environmental Protection Agency (EPA) reports that road transport contributed 83% of GHG emissions from the transportation sector, and the transportation sector accounts for 27% of all emissions in the USA (EPA 2014).

Sustainable development has become a global challenging issue. It is critical to use construction materials and apply intelligent design approaches that will help to maintain economic and environmental sustainability during infrastructure development and rehabilitation. Many local governments in the USA (e.g. City and County of Denver) have already developed policies or initiatives for cities to grow sustainably. A GHG reduction plan has been established by the City and County of Denver. The city supports growth patterns of public transportation, and develops more renewable energy sources to reduce GHG emissions. There is growing interest to quantify and compare the performance of pavements with alternative materials and designs because of the huge economic and environmental impacts of the roadway network. A good estimation on project-level life-cycle costs and environmental impacts is one of the important steps in the highway investment decision-making process. Life-cycle cost analysis (LCCA) can evaluate the economic impacts and life-cycle assessment (LCA) can estimate the environmental burden of the pavement system by investigating all phases of the pavement life cycle.

This article examines the current LCCA employed by the Colorado Department of Transportation (CDOT) in their pavement investment decision-making process, and proposes a regional LCA model to evaluate the GHG emissions from Colorado highway pavements. Both LCCA and LCA are used to quantify the economic and environmental impacts of a highway reconstruction project for alternative materials and designs. The limitations of current LCCA in decision-making are discussed. First, this article reviews the developments of pavement LCCA and LCA, summarizes the LCCA practice in CDOT and proposes a regional LCA model for Colorado pavements. Finally, the comparisons of LCCA and LCA results for a highway pavement construction project are discussed.

2. Pavement LCCA

2.1 Pavement LCCA development

The LCCA is a substantial component in an infrastructure management system, which evaluates the efficiency of investments (TRB 1985; U.S. Army 1986). LCCA includes agency costs (e.g. initial capital costs of construction, future costs of maintenance and rehabilitation, and residual value), as well as non-agency costs (e.g. user costs: occupancy time, operating costs, accidents, time delays due to maintenance and rehabilitation; nonuser costs: environmental pollution and neighbourhood disruptions).

LCCA for pavement was first discussed in the “Red Book” in the 1960s developed by the American Association of State Highway and Transportation Officials (AASHTO) (Wilde, Waalkes, and Harrison 2001). In early 1990s, pavement LCCA was included in the federal literature by employing several vehicle-operating-cost models (Zaniewski et al. 1982; Watanatada et al. 1987; Paterson and Attoh-Okine 1992; Uddin 1993). In 1995, Federal Highway Administration (FHWA) made LCCA compulsory for National Highway System projects costing more than $25 million, but this policy was annulled in 1998 under the Transportation Equity Act for the twenty-first century. However, FHWA and AASHTO are still providing guidance to states in developing their own LCCA procedures. Several bulletins were published, e.g. Life-Cycle Cost Analysis in Pavement Design interim technical bulletin (FHWA 1998), the Life-Cycle Cost Analysis Primer (FHWA 2002) and the Economic Analysis Primer (FHWA 2003). In addition, the “RealCost” LCCA software with a user manual has been provided by the FHWA, but use of this software is at the discretion of each state. According to Chan, Keoleian, and Gabler (2008), over 40 states in the USA implement LCCA during the pavement selection process. Although these states consider initial construction and future rehabilitation costs, only 40% incorporate user costs associated with road construction activities. In addition, none of these states consider nonuser social costs such as environmental damage (Chan, Keoleian, and Gabler 2008). The analysis period, pavement maintenance strategies and discount rates being used vary from state to state (Wilde, Waalkes, and Harrison 2001; Ozbay et al. 2004).

Two studies show that only instructional guidelines were provided in most states (Ozbay et al. 2004; ERES 2003). In addition, there are gaps existing between theoretical and actual LCCA applications (Gerke, Dewald, and Gerbrandt 1998; Carr 2000; Wilde, Waalkes, and Harrison 2001).

2.2 Pavement LCCA practice in Colorado

Pavement type selection and LCCA guidelines adopted by CDOT are included in its Pavement Design Manual (CDOT 2014). The FHWA Realcost programme is being used by CDOT for all new or reconstruction projects with more than $2 million initial pavement material and labour investment. Probabilistic LCCA comparing asphalt with concrete pavement is performed after the preliminary design for a typical section. If one pavement type has clearly better performance than the other (not less than 10% difference), the preliminary pavement type will be selected and a detailed pavement design will be performed. If not, the CDOT Pavement Type Selection Committee will evaluate other factors to determine the preliminary pavement type.

A 40-year analysis period and net present value (NPV) method are used in the LCCA (CDOT 2014). Planned rehabilitation with default input values is used in the LCCA to compare candidate strategies. The discount rate is calculated annually. In 2013, it was 2.8% with a standard deviation of 0.21%, and the 2014 discount rate is 2.6% with a standard deviation of 0.25%. The life-cycle cost factors include initial construction costs available from CDOT's cost data manual, maintenance costs included in the Pavement Design Manual, design costs, pavement construction engineering costs, traffic control, salvage value and user costs. The default values for average annual maintenance costs are $1270/lane-mile for hot-mixed asphalt (HMA) pavements, and $499/lane-mile for Portland cement concrete pavements (PCCPs) (CDOT 2014). An equation is provided in the Pavement Design Manual to estimate the salvage value. Realcost automates FHWA's work zone user cost calculation method, i.e. accounting the traffic delay caused by the construction activities. However, CDOT does not consider costs caused by crashes. Also nonuser costs are not included in the LCCA as well. Traffic data, e.g. average annual daily traffic (AADT) and single unit trucks/combination trucks as percentage of AADT, are used by pavement engineers to determine the pavement design parameters. The information is input into Realcost to calculate the user costs. The annual growth rate of traffic in Colorado defaults to a triangular distribution with minimum of 0.34%, maximum of 2.34% and most likely value of 1.34%. Values of time for passenger cars, single unit trucks and combination trucks are deterministic with $17/h, $35/h and $36.5/h, respectively. The values of time will be used in the environmental LCA to calculate the time of traffic delay due to the construction activities. Other input information in Realcost is detailed in the CDOT Pavement Design Manual. The LCCA is performed for a highway reconstruction project in this article, which compares HMA and PCCP alternatives.

3. Pavement environmental LCA

3.1 Pavement LCA

LCA quantifies the environmental impacts of all stages of a product, which include raw material acquisition, production manufacturing, transportation, installation, operation and maintenance, and ultimately, disposal, recycling and/or waste management. “Cradle-to-grave” LCA considers a product that ends life in the landfill. And “cradle-to-cradle” LCA considers a product that is recycled into new materials at the end of life (Liu et al. 2012).

LCA includes three models: process-sum LCA models, economy-wide LCA models and a hybrid LCA. Process-sum models track material and energy flows in all stages of a product. Economic data will be input in the economy-wide models, which is linked to energy use, emissions and toxic releases related with each industry sector. These models do not require arbitrary boundaries. One economy-wide model is the Economic Input–Output (EIO) model (CMDGI 2013). The hybrid LCA combines a process-sum model and economy-wide model.

The first pavement LCA literature appeared in the late 1990s. These studies have been published through a variety of sources, including industry organizations, peer-reviewed journals and government reports (Häkkinen and Mäkelä 1996; Horvath and Hendrickson 1998; Nisbet et al. 2001; Zapata and Gambatese 2005; Chan 2007; Zhang et al. 2010; Santero et al. 2011; Cass and Mukherjee 2011). Figure 1 identifies five distinct life cycle phases of the pavement in the literature: material production, construction, use, maintenance and end-of-life. Material production phase includes material manufacturing process from the extraction of raw materials (e.g. limestone) to the production of pavement materials (e.g. cement), including any necessary transportation. Construction phase includes onsite construction equipment and traffic delay due to work zone construction activities. After pavements are placed at the project locations, they interact with the environment through multiple pathways, e.g. vehicle rolling resistance and carbonation. These activities are included in the use phase. Maintenance phase includes the rehabilitation and reconstruction activities that occur during the life of a pavement. End-of-life phase can include demolition, disposal in a landfill, recycling processes and/or other activities that occur when the pavement is taken out of service.

Figure 1 Pavement LCA.

Several pavement LCA tools are available to the public, e.g. aspect (TRL 2011), BenRemod (Apul 2011), GHANGER (IRF 2011), GreenDOT (AASHTO 2010), PaLATE (Horvath 2004) and PE-2 (MTU 2011). Numerous LCA tools and models include pavements within their general scope (e.g. BEES, EIOLCA, SimaPro and Gabi).

As Santero, Masanet, and Horvath (2010) pointed out that

ideally, an LCA will examine each phase of the product life cycle across all relevant environmental impact categories in exhaustive detail. However, given time, data, and knowledge constraints, this process is very difficult for most products, including pavements. All LCAs are thus forced to simplify their scope and examine only those phases and processes that can be reasonably characterized under the study's constraints.

A regional deterministic environmental cradle-to-grave model was created as part of this work to evaluate the GHG emission associated with the Colorado highway pavements with different materials and design alternatives. Concrete is recyclable, but it is not common to recycle concrete pavement materials back to highway construction. Asphalt can be recycled up to certain percentage in highway application. However, additional GHG emissions and energy consumption are associated with the recycling process. In the future, a probabilistic cradle-to-cradle LCA model is needed to evaluate the environmental impacts of material recycling. The system boundaries of the regional cradle-to-grave model are discussed in the following section. A spreadsheet was designed to calculate the emissions. The environmental impact can act as an optional criterion for the selection of the preliminary pavement type.

3.2 Regional LCA model for Colorado pavements

A regional LCA, proposed for Colorado pavements, which can simulate the service life of new constructed pavements or pavement overlay systems, was performed in this study. The analysis period is 40 years. The functional units of the outputs used in this model were selected as kilogram carbon dioxide equivalent mass (kgCO₂e) per lane-km for GHG emission. There are six major gases that contribute to GHG, which are carbon dioxide, methane, nitrous oxide, hydrofluorocarbons, sulphur hexafluoride and perfluorocarbons. Different gases have varying impacts on global warming. Therefore, the impact of each gas is converted to the carbon dioxide equivalent mass (CO₂e) that would have the same climate impact. The LCA model includes the five life cycle phases of pavement identified in Figure 1. But in order to simplify the calculation of GHG emission, the model is divided into six modules: materials, construction, transportation, traffic congestion, usage and end-of-life. The material module consists of extraction and processing of raw materials. The construction module includes all construction activities, e.g. new construction, maintenance, rehabilitation and repairs. The GHG emission can be accounted from on-site construction equipment usage. The transportation module consists of transport of materials and equipment to and from the construction site, and transport of wastes to landfill. The traffic congestion module models the GHG emission due to traffic delay caused by construction and maintenance activities. The usage module calculates the additional fuel consumption and GHG emission due to the deteriorating pavements. The end-of-life module models demolition and landfill of the pavement materials.

The materials and transportation modules are modelled using data-sets from various sources. Reiner (2007) developed a regional concrete LCA model by analysing the primary materials, modes and costs of transport, processes for each life cycle phase for LCA of concrete flows in Colorado. Based on Reiner's regional LCA model, the emission factors of concrete materials except the cement are listed in Table 1. The authors quantified GHG emission of the ASTM C1157 Type GU Portland-limestone cement and ASTM C150 Type I/II cements produced in Colorado, based on the CO₂ Accounting and Reporting Standard for the Cement Industry (World Business Council for Sustainable Development 2005) developed by the Cement Sustainability Initiative of the World Business Council for Sustainability Development. The emission factors of Type GU cement and Type I/II cement are 0.819  and 0.873 kg CO₂e/kg, respectively. Comparing Type I/II cement, Type GU cement including 10% limestone emits less GHG. More than 100 miles of highway PCCP has been constructed with type GU cement in Colorado.

Table 1 Emission factors.

Material and construction process	Colorado study (kg CO2e/kg)	Sources for Colorado study
Aggregate/recycled concrete crushing	0.008	Reiner (2007)
Concrete mixing	0.019	Reiner (2007)
Cement	0.819/0.873	Holcim Type GU/Type I/II
Fly ash	− 0.018	Reiner 2007
Onsite equipment	0.0025	Zapata and Gambatese (2005)
Water	0.003	Reiner (2007)
Landfill of concrete	0.019	Reiner (2007)
Bitumen	0.19	Eurobitume (2011)
Truck transportation	0.0003	Reiner (2007)
Rail transportation	0.00003	Reiner (2007)

Fly ash is a by-product from coal-fired power plants. It can be recycled in concrete because of its pozzolanic, and in some cases, cementitious properties and does not contribute to direct emission of CO₂ in concrete. Fly ash has a negative emission factor due to the avoidance of landfill (Reiner 2007). Other materials (e.g. steel dowels, bitumen and adhesion agents) are made in other states or imported from other counties, so the industry averaged emission factors are included in the LCA model.

In the construction module, two methods are proposed to evaluate the GHG emission. The first method is to use the emission factor listed in Table 1. The alternative method is to estimate the operating time of the equipment during construction by using previous construction project documents of the CDOT. The US EPA NONROAD2008 model can be employed to calculate the fuel-related emissions.

The traffic delay considered in this study was caused only by the construction, maintenance, rehabilitation and repair activities. The congestion caused by crashes was not considered. The delays were calculated using Realcost software with input parameters such as AADT, work zone speed limits and lane capacity. Then, the traffic delays are coupled with the US EPA fuel economy guide and vehicle emissions to measure environmental impacts.

The usage phase is more complicated than other modules. Some pavement LCA models include the emission of the traffic during normal operation of the pavement. However, it is more reasonable to include the additional emission due to the deterioration of the pavements. There are three primary factors influencing fuel consumption and vehicle emissions: annual growth rate of the traffic, fuel economy and pavement deterioration. The annual growth rate of the traffic in Colorado has a triangular distribution. The most likely value, 1.34%, is used in this model. The fuel economy is described by Equation (1), which was adopted in a pavement LCA model developed by Zhang et al. (2010).

(1)

where FE_n is the fuel economy factor for nth year, FE_base is the baseline fuel economy factor and r is annual fuel economy improvement.

The baseline fuel economy factors were developed by the U.S. DOE (2004). The fuel economy is predicted to improve 1.5% per year for heavy-duty trucks due to improved design changes. Davis and Diegel (2002) estimated passenger cars and other light-duty vehicles to have an annual fuel improvement of 1%. The additional fuel consumption can be estimated by Equation (2), which was proposed by Epps et al. (1999), and adopted in the model developed by Zhang et al. (2010) as well.

(2)

where FCF is the fuel consumption factor (larger than 1.0) and IRI is the international roughness index.

The IRI affects the fuel economy and vehicle emissions. The final result is the difference between fuel consumed on an ideally smooth overlay and a rough pavement. According to Archonodo-Callao (Eurobitume 2011), the pavements have comfortable riding conditions with 3.5–4.5 m/km IRI. CDOT pavement inspection records provide the pavement deterioration trend of the concrete and HMA pavements.

In the end-of-life module, the LCA assumes that all materials are disposed in the landfill. The emissions from the landfill can be calculated using the emission factor listed in Table 1. Zhang et al. (2010) discussed that concrete pavement is not widely recycled due to the quality of the concrete decreasing with more than 20% recycled concrete materials. Around 80% of HMA pavement can be recycled into highway applications. However, special processing equipment and quality control procedures are needed to ensure the quality of the HMA with recycled asphalt materials. The environmental impacts of recycling procedures of PCCP and HMA should be included in the future LCA models.

4. Case study

LCCA and LCA were performed for CDOT Project IM C040-029. The project is to improve the capacity of interstate highway I-25 beginning on the north end of Colorado Springs at MP 149.3 to the north along I-25 to MP 154. The reconstructed highway consists of four lanes per direction, a 3.66 m (12-foot) wide inside shoulder and a 3.05 m (10-foot) wide outside shoulder. According to the preliminary soil survey, two alternative pavement sections were evaluated for this project:

• PCCP reconstruction consists of 33 cm (13 inches) of PCCP over 15.2 cm (6 inches) aggregate base course (ABC Class 6) placed on a minimum 61 cm (2 feet) of embankment soil with an AASHTO soil classification of A-1 or better. The PCCP has an initial design life of 30 years.

• HMA reconstruction consists of 20.3 cm (8 inches) HMA over 15.2 cm (6 inches) ABC (Class 6) placed on a minimum 61 cm (2 feet) of embankment with a minimum R-value of 60. The initial design life is 20 years.

The concrete thickness was designed according to the 1998 supplement to the AASHTO Guide for Design of Pavement Structures, Rigid Pavement Design. The HMA pavement was designed using the AASHTO computer program DARWin. The 81 kN (18-kip) design equivalent single-axle loads utilized in the PCCP and HMA designs were 44 million and 18 million, respectively, which were obtained from the Colorado Division of Transportation Development website and were based on 2010 published traffic volumes. The rehabilitation for the PCCP occurs at year 27. It consists of 0.5% full-depth PCCP slab replacement in the driving lanes, saw and seal joints, and full width diamond grinding. For the HMA alternative, rehabilitations occur at years 13, 26 and 39. The first rehabilitation uses polymerized stone-mixed asphalt (SMA) wearing course for repair. The second and third rehabilitations consist of 5 cm (2 inches) mill and fill treatments with polymerized (SMA). Regardless of the alternate pavement type, two lanes of through traffic per direction should be maintained throughout the initial construction and rehabilitation period.

All materials consumed in the two alternatives are summarized in Table 2. And the material proportions of concrete and HMA are listed in Table 3. The above-mentioned information is used for the LCCA and environmental LCA.

Table 2 Total material consumption.

	Quantity	Unit	Quantity	Unit
PCCP alternative
PCCP	70,292	yd^3	53,742	M^3
ABC Class 6	33,724	t	30,594	Mt
Embankment	68,593	yd^3	52,443	M^3
HMA alternative
SMA	22,483	t	20,396	Mt
HMA	67,448	t	61,188	Mt
ABC Class 6	33,724	t	30,594	Mt
Embankment	68,593	yd^3	52,443	M^3

Table 3 Material proportions.

PCCP	lb/cy	kg/m^3	Asphalt	% per unit volume
Cement	448	264	Bitumen	7
Fly ash	112	66	Gravel	71
Rock	1780	1050	Sand	17
Sand	1310	773	Limestone	5
Water	237	140	–	–

5. LCCA and LCA results and discussion

5.1 LCCA

A probabilistic LCCA was completed using the NPV economic analysis over a 40-year period. The agency costs include materials, preliminary engineering, construction and traffic control. In addition to the material costs, preliminary engineering for each initial construction alternative and rehabilitation was assumed at 10%, construction was assumed at 17.45%, and traffic control at 15%. The Realcost outputs are summarized in Table 4.

Table 4 LCCA.

	HMA alternative		PCCP alternative
Statistics	Agency cost$ thousands	User cost$ thousands	Agency cost$ thousands	User cost$ thousands
Minimum	12,715.71	84.64	12,450.81	41.94
Maximum	20,493.24	505.76	20,961.32	1,053.36
Mean	16,613.59	245.55	16,873.32	674.80
Median	16,632.92	222.88	16,873.48	730.15
Standard deviation	1,390.11	73.00	1,752.60	202.40
Percentile (5%)	14,358.70	151.29	14,010.81	255.01
Percentile (10%)	14,865.76	169.40	14,591.88	353.95
Percentile (75%)	17,578.75	291.46	18,147.20	809.77
Percentile (95%)	18,897.43	386.30	19,783.69	923.48

Table 4 shows the probabilistic LCCA with 95% confidence level is 7.4% in favour of the HMA over PCCP. Because the LCCA indicates the difference is less than 10%, the designs are considered to be equivalent. A Pavement Type Selection Committee was recommended, according the CDOT Pavement Design Manual.

5.2 Environmental LCA

In the LCA, ASTM 1157 Type GU and ASTM 150 Type I/II cements are used to produce the PCCP. The differences of the environmental impacts from the two PCCPs lie in the materials module only. The GHG emission of the 4.7 miles (7.6 km) highway from the materials module can be calculated using the information in Tables 123. The transportation module is closely related to the material and end-of-life module. All materials and wastes in this project are distributed by trucks in Colorado. Table 5 shows the estimated transportation distances for different materials and wastes. The information in Tables 1 and 5 can be used to calculate the emissions from transportation modules.

Table 5 Transportation distances.

Materials	Distance (miles)	Truck (km)
Cement plant to site	100	161
Sand source to ready mix	50	80
Gravel source to site	50	80
Water source to site	3	5
Fly ash source to site	100	161
Bitumen source to site	50	80
Site to landfill	50	80

In the construction module, the operating time of the equipment is estimated using previous project documents, indicated in Table 6. Combined with the US EPA NONROAD2008 model of diesel engine emissions for Colorado, the fuel usage can be estimated. Multiplying the fuel usage by the diesel emission factor available from the US EPA(4), the GHG emissions can be calculated.

Table 6 Total estimated equipment usage during construction.

Equipment	Power (kW)	PCCP (h)	HMA (h)
Crawler-mounted hydraulic excavator	320	410	410
Air compressor	260	205	205
Dumper	17	538	461
Hydraulic hammer	75	256	205
Motor grader	123	51	51
Water truck	335	102	102
Vacuum truck	132	102	102
Wheeled front-end loader	175	666	615
Signal boards	4	1000	1215
Concrete paver	186	1059	0
Concrete truck	223	1059	0
Resonant breaker	447	0	320
Asphalt paver	150	90	1450
Asphalt roller	93	26	504
Asphalt truck	223	90	1355

The traffic delay considered in the LCA is caused only by the construction, maintenance and rehabilitation activities. Table 7 shows the percentages of various types of vehicles and values of user time. Based on the information and user cost calculated from Realcost 2.5 software, the time delayed can be calculated. The emissions during the delayed time can be estimated according to the US EPA (4).

Table 7 Percentages of vehicles and values of user time.

	Percentage	Value of user time
Passenger cars	90.50	$17
Single unit trucks	3.60	$35
Combination trucks	5.90	$36.50

According to the proposed method discussed in the previous section, the GHG emission from additional fuel consumed due to the deteriorated pavement and from the landfill can be calculated. The emissions from the six modules are summarized in Table 8. In the 40 years of analysis period, the PCCP has more GHG emission than the HMA alternative in the module of materials. But comparing the overall emissions, the GHG emission from PCCP is 26% less than that of the HMA. Among the six modules, the emissions from the use phase due to the deterioration of the pavements control the overall performance. Since HMA pavement deteriorates quicker than the PCCP, it is reasonable to conclude that the PCCP has better environmental performance than the HMA pavement. The GHG emissions from PCCP and HMA per lane-km are 4.28 thousand metric tons for Type GU PCCP, 4.31 thousand metric tons for Type I/II PCCP, and 5.84 thousand metric tons for HMA, respectively, in the 40-year analysis period.

Table 8 GHG emissions.

	PCCP Type GU		PCCP Type I/II		HMA
Sectors	kg CO2e	kg CO2e/lane-km	kg CO2e	kg CO2e/lane-km	kg CO2e	kg CO2e/lane-km
Materials	1.38E+07	2.29E+05	1.54E+07	2.56E+05	4.32E+06	7.18E+04
Transportation	3.27E+06	5.44E+04	3.27E+06	5.44E+04	6.46E+07	1.07E+06
Construction	3.10E+05	5.15E+03	3.10E+05	5.15E+03	3.57E+05	5.93E+03
Traffic delay	2.76E+06	4.59E+04	2.76E+06	4.59E+04	1.46E+06	2.43E+04
Use phase	2.35E+08	3.91E+06	2.35E+08	3.91E+06	2.78E+08	4.62E+06
Landfill	2.35E+06	3.91E+04	2.35E+06	3.91E+04	2.71E+06	4.51E+04
Total	2.57E+08	4.28E+06	2.59E+08	4.31E+06	3.51E+08	5.84E+06

5.3 Comparison of LCCA and LCA

The LCCA indicates that the two alternatives are equivalent, but it is 7.4% in favour of the HMA alternative. In the LCA, PCCP emits more GHG from materials and traffic delay modules than the HMA. However, the overall environmental performance of PCCP is superior over the HMA because the emission from the use phase controls the overall performance. Comparing the LCCA and LCA, the user cost caused by the additional fuel consumption due to the pavement deterioration is not considered in the current LCCA. Because the PCCP has a better environmental performance in the use phase, it is expected that the PCCP has a lower user cost than the HMA.

6. Conclusion

The roadway network is one of the most important infrastructures in the USA. Materials with lower environmental impacts and intelligent engineering designs are required to support sustainable development. One substantial step to improve the pavement investment decision-making process is to estimate the life-cycle cost and environmental impacts of various alternatives. The developments of pavement LCCA and environmental LCA were summarized in this article. The current LCCA practice used by CDOT was reviewed and a regional pavement LCA model was developed in this study. Both LCCA and LCA were used to analyse a CDOT highway reconstruction project. The economic and environmental performances of the two alternatives were evaluated and compared. In the LCCA, the HMA indicates a slightly better performance than the PCCP. However, because the HMA pavement has a higher deterioration rate than the PCCP pavement, the HMA pavement consumes more fuel of vehicles than the PCCP pavement in the 40-year analysis period. But the LCCA does not consider the additional user cost in the usage phase of the pavement. The environmental LCA indicates that the emission from the usage phase controls the performance of the pavement, which is much higher than other sectors. Overall, the PCCP has 26% less GHG emissions than the HMA alternative over the 40-year analysis period. The user cost in the usage phase may control the economic performance of the pavement as well if it is considered in the LCCA. The PCCP pavement outperforms HMA pavement in this case study. This study also shows that both LCCA and LCA provide support for a management decision, but do not alone dictate a decision. In addition to the comparison of economic and environmental performances of different alternatives, soil characteristics, traffic condition, climate and construction considerations should be considered in any pavement design.

Acknowledgements

The authors would like to thank Dr Stephan A. Durham, College of Engineering, University of Georgia, Mr Don A. Clem from Portland Cement Association, and various personnel from Colorado Department of Transportation for their assistance with this study.