Final Report: Computer-Aided Hybrid Models for Environmental and Economic Life-Cycle Assessment

EPA Grant Number: R829597
Title: Computer-Aided Hybrid Models for Environmental and Economic Life-Cycle Assessment
Investigators: Horvath, Arpad , Hendrickson, Chris , Eyerer, Peter
Institution: University of California - Berkeley , University of Stuttgart , Carnegie Mellon University
Current Institution: University of California - Berkeley , Carnegie Mellon University , University of Stuttgart
EPA Project Officer: Hahn, Intaek
Project Period: January 1, 2002 through December 31, 2004
Project Amount: $325,000
RFA: Technology for a Sustainable Environment (2001) RFA Text |  Recipients Lists
Research Category: Sustainability , Pollution Prevention/Sustainable Development

Objective:

The overall goal of this research project was to develop hybrid models that will overcome the major limitations of the two life-cycle assessment (LCA) approaches practiced currently: one based on detailed process model descriptions, corresponding emissions, and wastes (process-based LCA), and the other based on economic input-output (EIO) data and publicly available resource consumption and environmental discharge data (EIO-LCA). Although both approaches have advantages, both have major limitations as well. We demonstrated the utility and comprehensiveness of hybrid models combining both LCA approaches by applying them to life-cycle studies from different sectors of the economy. The specific objectives of this research project were to: (1) determine how hybrid LCA models lead to more comprehensive and less uncertain results for a given application; (2) find out how hybrid LCA models are useful to what level of decisionmaking; and (3) determine the accuracy and comprehensiveness of hybrid LCA models against the stand-alone process-based and EIO-LCA models.

Summary/Accomplishments (Outputs/Outcomes):

We have focused on identifying the ways of integrating the process-based and the EIO-LCA models with respect to: (1) including detailed process-level environmental data as well as economy-wide (supply chain) environmental impacts; (2) having environmental and economic information about the major products and processes in the economy; and (3) quantifying a wide range of environmental data.

We have compared comprehensively the strengths and weaknesses of the two LCA models and identified gaps where the hybrid model could prove superior to either stand-alone model. We also have reviewed the coverage of environmental emission and waste factors of the two LCA tools that were available to us: GaBi (process-based LCA) and EIO-LCA (input-output analysis-based LCA).

We have prepared both GaBi and EIO-LCA for hybrid analyses by exchanging relevant information that was missing from either tool. For example, many sectors from EIO-LCA have been identified that currently do not exist in GaBi and that could be added to GaBi with the understanding that they represent U.S. conditions and data. Further, EIO-LCA could benefit from more detailed coverage of plastics that currently is missing but could be transferred from GaBi or another process-based LCA project. In terms of the coverage of environmental emissions and impacts, there were similarities found between the tools, but often the definitions of emissions or waste categories were different. We found it desirable to increase the common set of environmental aspects.

Beyond the modeling of hybrid LCA, we have expanded much effort in operationalizing and applying hybrid LCA. Publication of papers and completion of Ph.D. dissertations were the best way to describe what problems we tackled in this research with the new modeling method.

Comparing Electricity Generation Technologies and Impacts

Our first paper out of this research was a groundbreaking one in that it was the first peer-reviewed journal paper that compared comprehensively five electricity generation technologies: three renewable (hydro, wind, solar photovoltaics) and two nonrenewable (coal and natural gas) options in the construction, operation, maintenance, and end-of-life (EOL) phases.

As demand for electricity increases, investments into new generation capacity from renewable and nonrenewable sources should include assessment of climate change consequences not just of the operational phase of the power plants but construction effects as well. In this research, we have quantified the global warming effect (GWE) associated with construction and operation of comparable hydroelectric, wind, solar, coal, and natural gas power plants for four periods after construction. The assessment includes greenhouse gas emissions from construction, burning of fuels, flooded biomass decay in the reservoir, loss of net ecosystem production, as well as land use. The results indicate that a wind farm and a hydroelectric power plant in an arid zone (such as the Glen Canyon in the Upper Colorado River Basin) have lower GWE than other power plants. For the Glen Canyon power plant, the upgrade 20 years after the beginning of operation increased power capacity by 39 percent but resulted in a mere 1 percent of the CO2 emissions from the initial construction and came with no additional emissions from the reservoir, which accounts for the majority of the GWE.

The result of this project was a paper (Pacca and Horvath, 2002) that has quantified not just the global warming gas emissions from the smokestacks of coal- and natural gas-fired power plants, but has estimated the emissions from hydroelectric plants in their use phase (i.e., the effects of the reservoir) as well, and has quantified comprehensively the emissions from the construction of five different power plant types.

This project also looked at how existing LCA models consider the issue of scale as related to estimating impacts of producing electricity in the United States. Most models use national-level averages by internalizing the national grid mix (e.g., 50% coal-fired). There are large regional variations, however, in the mix of electricity consumed. The paper by Marriott (2005) estimated mixes of consumed electricity in each of the 50 states and noted large variations of impacts between using national averages and actual state consumption. The output of this work was the creation of improved estimates of impacts from electricity production for use in EIO-LCA.

Energy and Air Emissions Assessment of Telecommuting/Telework

Telecommuting, or telework, has emerged as a popular employment benefit that also is thought to save energy and air emissions. More and more employees telecommute or telework worldwide.

We have worked on quantifying and comparing the energy demand and air emissions from nontelework and telework scenarios involving transportation, heating, cooling, lighting, electronic, and electrical equipment use both at the company and the home office. Monte Carlo simulation was used to perform probabilistic and sensitivity analyses over a set of scenarios based on literature, national surveys, and estimated data.

Telework is becoming a practice that calls for careful quantification of the private and social costs and benefits. We have found that telework may not affect equally the emissions of all types of pollutants: it may decrease CO and NOx emissions, but increase CH4, CO2, N2O, coarse particulate matter (PM10), and SO2 emissions under the assumed conditions. Therefore, the scope and goal of telework programs must be defined early in the implementation process. Transportation-related impacts could be reduced as a result of telework; however, home-related impacts caused by an employee spending additional time at home could offset these reductions. Company office-related impacts may not be reduced unless the office space is shared with other employees during telework days, or eliminated entirely. The success of telework programs is found to depend mainly on commuting patterns, induced energy usage, and characteristics of office and home space use. Under the assumed scenarios, more CO2, SO2, CH4, Hg (except California and Illinois), N2O, and PM10 emissions would be generated by teleworking than by nonteleworking employees in five states with high telework potential (California, New York, Texas, Georgia, and Illinois). Further, we have found that the shifts in the mode of employee transportation, the transportation technology used, and occupancy rate changes can alter significantly the anticipated environmental costs and benefits of telework programs. There is a distinct correlation between telecommuting frequency, the vehicle miles traveled, and the nontelework to telework impacts ratio that should be considered when planning for and evaluating telework programs. Costs identified in literature such as increased energy usage at home, remote access, equipment, and program set-up costs do not seem to overcast the general expectation that telework could increase company profits. Individuals’ utility costs may increase, but commuting and vehicle maintenance costs, and other daily expenses (e.g., dry cleaning) may decrease, ensuring a favorable bottom line.

In collaboration with Eric Williams of the United Nations University in Tokyo, Japan (Matthews, 2005), we also did some system-wide energy consumption comparisons of telework strategies in the United States and Japan. As mentioned above, telework strategies should be planned carefully. Small-scale projects that do not reduce significantly commercial space needs may have rebound effects that end up using more energy and resources because of offsets in the energy-intensive residential sector.

Environmental Assessment of Residential and Commercial Buildings

Residential and commercial buildings are thought to be significant sources of energy use and emissions in, but quantitative assessments of all of the phases of the service life of these buildings have been missing. To enable environmentally conscious design and management, this research has completed LCAs of newly constructed European and U.S. buildings from materials production through construction, use, and maintenance to EOL treatment.

For commercial buildings, we have quantified the significant environmental aspects of a new high-end office building during 50 years of service life. A comprehensive environmental LCA, including data quality assessment, has been conducted to provide detailed information for establishing the causal connection between the different life-cycle elements and potential environmental impacts. The results showed that most of the impacts are associated with electricity use and building materials manufacturing, in particular electricity used in lighting, HVAC systems, and outlets; heat conduction through the structures; manufacturing and maintenance of steel; manufacturing of concrete and paint; water use and wastewater generation; and office waste management. Construction and demolition were found to have relatively insignificant impacts. The identified most significant aspects are quite predominant: 7 percent of all counted aspects cover more than 50 percent of the life-cycle impacts. Practical applications of the study’s results could be in environmentally conscious design and management of office buildings.

For residential buildings, we estimated the building resource requirements, electricity and energy used, greenhouse gas releases, hazardous waste generated, and toxic air releases for the construction, usage, and demolition of typical U.S. residences in 1997. Within the three phases, usage (54% of economic activity) is the largest consumer of electricity (95%) and energy (93%) and the largest emitter of greenhouse gases (92%), whereas the construction phase (46% of economic activity) is the largest air toxics emitter (57%) and contributes 51 percent of hazardous waste. The disposal phase contribution is negligible in all of these categories. From the standpoint of the entire U.S. economy, residential buildings account for 5.3 percent of the Gross Domestic Product, 38 percent of electricity consumption, 26 percent of energy consumption, 24 percent of greenhouse gas emissions, 26 percent of hazardous waste, and 12 percent of toxic air emissions. We commented on possible remedial actions, including some current public policies, to address environmental impacts.

For both types of buildings, the significant environmental aspects indicate the dominance of the use phase in the quantified environmental categories but draw attention to the importance of embedded materials and expected maintenance investments throughout the assumed 50-year service life, especially for particulate matter emissions. The relevance of the materials, construction, maintenance, and EOL phases relative to the use of buildings is expected to increase considerably as functional obsolescence of office buildings becomes more rapid and complete reconstruction and reconfiguration become more frequent. By quantifying the energy use and environmental emissions of each life-cycle phase in more detail, the elements that cause significant emissions can be identified and targeted for improvement. This information will benefit both industry practitioners and researchers so that they can focus their environmental efforts where they can have the most impact.

Life-Cycle Assessment of Water Provision

In many parts of the world, water availability is diminishing because of scarce sources, growing population, inefficient use, and pollution. The environmental implications of water supply options, both traditional (groundwater or surface water) or alternative sources (desalination, recycling), are not well understood because of a lack of LCA studies. As readily available water is depleted, subsequent alternatives likely will have higher energy and resource requirements and, therefore, environmental impacts. To develop a more environmentally responsible and sustainable water supply system, these environmental implications should be incorporated into planning decisions. Accounting for energy and environmental effects in water planning requires LCA.

In this research, LCA was used to compare three supply alternatives: importing, recycling, and desalinating water. Water conservation using water-efficient fixtures also was evaluated for comparison. Energy use and environmental emissions were reported for the water supply alternatives, for life-cycle phases, and for water supply functions (i.e., supply, treatment, and distribution). The Water-Energy Sustainability Tool has been developed to evaluate water planning decisions with a life-cycle perspective. As case studies, the tool was used to evaluate the systems of two California water utilities, the Marin Municipal Water District and the City of Oceanside Water Department. The results showed that for both utilities desalination was the most environmentally detrimental primarily because of the energy-intensity of the treatment process. The recycled and imported water results were less conclusive because they were affected more by the distance to the water source, topography, treatment process used, and other issues. For all alternatives in both case study systems, energy consumed by system operation dominated the results. The results from this study can inform future water supply planning.

Environmental Assessment of Wireless Technologies

Wireless information technologies are providing new ways to communicate and are one of several information and communication technologies touted as an opportunity to reducing society’s overall environmental impacts. Rigorous system-wide environmental impact comparisons of these technologies to the traditional applications they may replace, however, have been initiated only recently, and the results have been mixed. In this research, the environmental effects of two applications of wireless technologies have been compared to those of conventional technologies for which they can substitute. First, reading newspaper content on a personal digital assistant (PDA) is compared to the traditional way of reading a newspaper. Second, wireless teleconferencing is compared to business travel. The results show that for both comparisons wireless technologies create lower environmental impacts. Compared to reading a newspaper, receiving the news on a PDA wirelessly results in the release of 32-140 times less CO2, several orders of magnitude less NOx and SOx, and the use of 26-67 times less water. Wireless teleconferencing results in one to three orders of magnitude lower CO2, NOx, and SO2 emissions than business travel.

This research has been featured in Science News, Globe and Mail (Toronto), Environmental News Network, and Great Lakes Radio Consortium.

Air Emissions Inventory of Transportation and Logistics

Environmental awareness increasingly is important to society, government, and industry, and there is a strong demand for sustainable development practices. The importance of supply chain management is critical as it characterizes and influences the life-cycles of all products.

Within the major logistics trends, outsourcing has a significant potential to increase sustainability in the supply chain as third-party logistics providers (3PLs) focus on improving resource utilization and making processes more efficient. Their motivation, however, is largely economic, and an environmental perspective is rarely seen in 3PLs. As consumers demand greener alternatives and environmental regulatory measures subsequently are implemented, 3PLs will have to become environmentally and socially more aware to develop sustainability goals. This study compared two scenarios using LCA: one where logistics functions were handled in-house, and an alternative scenario where such functions were outsourced to a 3PL. The impacts of logistics outsourcing on energy utilization, global warming potential, and fatalities were first quantified in the supply chain of an automobile. Even though vehicle operation, responsible for most of the impacts considered, was outside the domain of logistics functions, logistics outsourcing nonetheless has the potential to reduce energy use and global warming potential by 0.4-2 percent and fatalities by 0.8-3.3 percent throughout the entire life-cycle of a typical automobile. Road and air transportation are found to account for most of the impacts in all selected metrics. Analyzing logistics outsourcing in the other sectors of the U.S. economy revealed the same trend as observed in the supply chain of an automobile.

The other study we undertook analyzed the inventory of air emissions (CO2, NOx, PM10, CO, SO2, and Pb) associated with the transportation of goods by road, rail, and air in the United States. It includes the manufacturing, use, maintenance, and EOL phases of vehicles; construction, operation, maintenance, and EOL of transportation infrastructure; and oil exploration, refining, and fuel distribution.

This was the first comprehensive study to compare road, rail, and air transportation of goods in the United States in terms of life-cycle emissions. It benefits those performing environmental research on transportation, including the internalization of social costs. Governments will receive input for the development of transportation policies that support more effective decisionmaking. As consumers become more environmentally aware, and governments start promoting regulations to internalize environmental externalities into transportation prices, environmental criteria will play a bigger role in transportation decisions by manufacturers and carriers. Additionally, more comprehensive information on life-cycle air emissions is useful to support the development of more accurate LCAs of diverse products.

Hybrid LCA was the methodology used for the comparison. All the components were summed by a common functional unit of tons of air pollutant per ton-mile of freight activity. Even though results reflected a baseline scenario, the model also enabled the differentiation of results based on alternative vehicle configurations (e.g., vehicle sizes), business practices (e.g., vehicle utilization, empty backhaulage), and geography (e.g., energy mix, road grade, rail track grade).

Depending on the pollutant, rail transportation rates were 40-90 percent better than road transportation. Air transportation scored the lowest in terms of environmental performance, being from 2 to 16 times worse than road transportation. The ranking, however, can change by accounting for intramodal variations (e.g., small trucks versus large trucks) and uncertainty. This is the case for CO and SO2 where road and rail overlap, as well as for PM10 and Pb where air and road overlap. The ranking of modes also can change for some pollutants depending whether the analysis accounts for total emissions or only fuel combustion emissions. This was the case of CO and SO2 emissions, where rail scored worse than road transportation if only fuel combustion is included in the analysis.

Results confirm that total emissions are significantly underestimated if only fuel combustion emissions are accounted for. Fuel combustion accounts for 70-90 percent of total emissions of CO2. NOx emissions tend to be more concentrated in the fuel combustion phase (70-98%). Infrastructure is the predominant phase for PM10 emissions caused by construction processes, especially for road and air transportation. Because of regulations of CO emissions from road transportation, truck manufacturing dominates CO emissions from road transportation. Fuel combustion is still the predominant phase for CO emissions from rail transportation caused by lack of emission standards. Airport operations rate high in CO emissions because of the operation of ground support equipment. In terms of SO2 emissions, road and air infrastructure are the predominant phases caused by road construction processes, and ground support equipment, respectively. Fuel combustion is representative in rail transportation because of a lack of fuel sulfur reduction standards. Because of the use of unleaded fuel, vehicle manufacturing is responsible for the majority of lead emissions for all three modes. Differences between tailpipe emissions and total system-wide emissions can range from 2 percent for rail transportation’s NOx emissions to an almost 23-fold difference for air transportation’s PM10 emissions. Therefore, it is of paramount importance to consider infrastructure, vehicle manufacturing, and precombustion processes, whose life-cycle share is likely to increase as new tailpipe emission standards are enforced.

Fuel combustion emission factors, however, are still the most important parameter influencing total emissions of CO2 and NOx. Mode operational efficiency is of paramount importance to improve the environmental performance of freight transportation, and it can be reflected in improvements in equipment utilization and reductions in empty backhauls. Air emissions do not depend significantly on geographic location. Finally, there are significant differences between the environmental performance of freight transportation in the United States and Europe.

Transportation policy has focused traditionally on regulatory measures to curb fuel combustion emissions. Although this is effective for NOx emissions, which fall heavily on the fuel combustion phase, it lags behind for PM10, SO2, and CO emissions, as significant shares are released from infrastructure, vehicle, and precombustion phases. Environmental regulations applied to these phases also are necessary. Driven by economic incentives (e.g., high oil prices and lower margins), the transportation industry also has worked towards improving fuel efficiency, thus lowering emissions per-ton-mile. Governmental intervention also can impose more strict standards to the freight transportation industry. Even though environmental taxes are unpopular in the transportation industry, they are arguably one of the most effective ways to internalize environmental externalities in transportation prices and adjust industry behavior to take environment costs into account in their mode choice decisions. Other policies include additional user fees to transportation modes that have in higher emissions per-ton-mile (e.g., air transportation), as well as giving incentives to modes that are more environmentally friendly (e.g., rail transportation).

Parameters with the highest uncertainty (e.g., infrastructure and vehicle lifetime) have relatively low impacts on final results, assuring the model’s robustness. Uncertainty can be significantly reduced by analyzing a specific route (e.g., no empty backhaulage) where vehicle utilization is known. Further studies are necessary to address inland water shipping, as well as additional air emissions and other environmental effects. An impact assessment also is necessary to evaluate the impacts of the emissions inventory assembled in this study, which should be followed by an economic valuation of environmental impacts.


Journal Articles on this Report : 16 Displayed | Download in RIS Format

Other project views: All 38 publications 18 publications in selected types All 16 journal articles
Type Citation Project Document Sources
Journal Article Facanha C, Horvath A. Environmental assessment of logistics outsourcing. Journal of Management in Engineering 2005;21(1):27-37. R829597 (Final)
  • Abstract: ACS-Abstract
    Exit
  • Journal Article Facanha C, Horvath A. Environmental assessment of freight transportation in the US. International Journal of Life Cycle Assessment 2006;11(4):229-239. R829597 (Final)
  • Abstract: Springer-Abstract
    Exit
  • Journal Article Facanha C, Horvath A. Evaluation of life-cycle air emission factors of freight transportation. Environmental Science & Technology 2007;41(20):7138-7144. R829597 (Final)
  • Abstract from PubMed
  • Full-text: ACS-Full Text HTML
    Exit
  • Abstract: ACS-Abstract
    Exit
  • Other: ACS-Full Text PDF
    Exit
  • Journal Article Guggemos AA, Horvath A. Comparison of environmental effects of steel-and concrete-framed buildings. Journal of Infrastructure Systems 2005;11(2):93-101. R829597 (Final)
  • Abstract: ASCE-Abstract
    Exit
  • Journal Article Guggemos AA, Horvath A. Decision-support tool for assessing the environmental effects of constructing commercial buildings. Journal of Architectural Engineering 2006;12(4):187-195. R829597 (Final)
  • Abstract: ASCE-Abstract
    Exit
  • Journal Article Junnila S, Horvath A, Guggemos AA. Life-cycle assessment of office buildings in Europe and the United States. Journal of Infrastructure Systems 2006;12(1):10-17. R829597 (Final)
  • Abstract: ASCE-Abstract
    Exit
  • Journal Article Kitou E, Horvath A. Energy-related emissions from telework. Environmental Science & Technology 2003;37(16):3467-3475. R829597 (Final)
  • Abstract from PubMed
  • Full-text: ACS-Full Text HTML
    Exit
  • Abstract: ACS-Abstract
    Exit
  • Other: ACS-Full Text PDF
    Exit
  • Journal Article Kitou E, Horvath A. Transportation choices and air pollution effects of telework. Journal of Infrastructure Systems 2006;12(2):121-134. R829597 (Final)
  • Abstract: ASCE-Abstract
    Exit
  • Journal Article Marriott J, Matthews HS. Environmental effects of interstate power trading on electricity consumption mixes. Environmental Science & Technology 2005;39(22):8584-8590. R829597 (Final)
  • Full-text: ACS-Full Text HTML
    Exit
  • Abstract: ACS-Abstract
    Exit
  • Other: ACS-Full Text PDF
    Exit
  • Journal Article Matthews HS, Hendrickson CT. The economic and environmental implications of centralized stock keeping. Journal of Industrial Ecology 2002;6(2):71-81. R829597 (Final)
    R826740 (Final)
  • Full-text: Wiley-Full Text PDF
    Exit
  • Abstract: Wiley-Abstract
    Exit
  • Journal Article Matthews HS, Williams E. Telework adoption and energy use in building and transport sectors in the United States and Japan. Journal of Infrastructure Systems 2005;11(1):21-30. R829597 (Final)
  • Full-text: Semantics Scholar-Full Text PDF
    Exit
  • Abstract: ASCE-Abstract
    Exit
  • Journal Article Ochoa L, Hendrickson C, Matthews HS. Economic input-output life-cycle assessment of U.S. residential buildings. Journal of Infrastructure Systems 2002;8(4):132-138. R829597 (Final)
  • Full-text: Academia-Full Text HTML
    Exit
  • Abstract: ASCE-Abstract
    Exit
  • Journal Article Pacca S, Horvath A. Greenhouse gas emissions from building and operating electric power plants in the upper Colorado River Basin. Environmental Science & Technology 2002;36(14):3194-3200. R829597 (2002)
    R829597 (Final)
  • Abstract from PubMed
  • Full-text: ACS-Full Text HTML
    Exit
  • Abstract: ACS-Abstract
    Exit
  • Other: ACS-Full Text PDF
    Exit
  • Journal Article Stokes J, Horvath A. Life cycle energy assessment of alternative water supply systems. The International Journal of Life Cycle Assessment 2006;11(5):335-343. R829597 (Final)
  • Abstract: Springerlink Abstract
    Exit
  • Journal Article Suh S, Lenzen M, Treloar GJ, Hondo H, Horvath A, Huppes G, Jolliet O, Klann U, Krewitt W, Moriquchi Y, Munksgaard J, Norris G. System boundary selection in life-cycle inventories using hybrid approaches. Environmental Science & Technology 2004;38(3):657-664. R829597 (2002)
    R829597 (Final)
  • Abstract from PubMed
  • Full-text: ACS-Full Text HTML
    Exit
  • Abstract: ACS-Abstract
    Exit
  • Other: ACS-Full Text PDF
    Exit
  • Journal Article Toffel MW, Horvath A. Environmental implications of wireless technologies: news delivery and business meetings. Environmental Science & Technology 2004;38(11):2961-2970. R829597 (Final)
  • Abstract from PubMed
  • Full-text: ACS-Full Text HTML
    Exit
  • Abstract: ACS-Abstract
    Exit
  • Other: ACS-Full Text PDF
    Exit
  • Supplemental Keywords:

    life-cycle analysis, decision-making, pollution prevention, industrial ecology, sustainable development, clean manufacturing, environmental cost analysis, green design, economics, clean technologies,, RFA, Scientific Discipline, Air, Sustainable Industry/Business, Sustainable Environment, climate change, Air Pollution Effects, Economics, Technology for Sustainable Environment, Economics and Business, Environmental Engineering, Atmosphere, computational simulations, environmental monitoring, life cycle analysis, cleaner production, clean technologies, green design, life cycle inventory, computer models, environmental sustainability, computer science, industrial ecology, clean manufacturing, computer generated alternatives, pollution prevention design, life cycle assessment, pollution prevention, environmental cost analysis

    Progress and Final Reports:

    Original Abstract
  • 2002 Progress Report
  • 2003