Article pubs.acs.org/est
Functional Unit, Technological Dynamics, and Scaling Properties for the Life Cycle Energy of Residences Stephane Frijia,† Subhrajit Guhathakurta,‡ and Eric Williams*,§ †
Greater Phoenix Economic Council, Phoenix, Arizona, United States School of City & Regional Planning, Georgia Institute of Technology, Atlanta, Georgia, United States § Golisano Institute of Sustainability, Rochester Institute of Technology, Rochester, New York, United States ‡
S Supporting Information *
ABSTRACT: Prior LCA studies take the operational phase to include all energy use within a residence, implying a functional unit of all household activities, but then exclude related supply chains such as production of food, appliances, and household chemicals. We argue that bounding the functional unit to provision of a climate controlled space better focuses the LCA on the building, rather than activities that occur within a building. The second issue explored in this article is how technological change in the operational phase affects life cycle energy. Heating and cooling equipment is replaced at least several times over the lifetime of a residence; improved efficiency of newer equipment affects life cycle energy use. The third objective is to construct parametric models to describe LCA results for a family of related products. We explore these three issues through a case study of energy use of residences: one-story and two-story detached homes, 1,500−3,500 square feet in area, located in Phoenix, Arizona, built in 2002 and retired in 2051. With a restricted functional unit and accounting for technological progress, approximately 30% of a building’s life cycle energy can be attributed to materials and construction, compared to 0.4−11% in previous studies.
■
Functional Units for Buildings. The definition of the functional unit is fundamental to any life cycle assessment. The functional unit is the unit of functionality associated with a product or service in question.1 To illustrate the concept, a functional unit to compare light bulb technologies could be chosen as 10,000 h of 1,800 lm light. The reference flow is the associated product/service systems needed to deliver the functional unit, e.g. one 23 W compact fluorescent light bulb in this example. We argue that for theoretical and practical reasons there is a need to revisit the definition of functional unit and reference flows for buildings. Prior studies analyze a reference flow of embodied materials and construction combined with entire operational energy within a building.5,15,16 The functional unit for such a reference flow is unclear. To clarify this statement, Table 1 shows functions associated with residences and how building systems and other supply chains connect with these services. Depending on the purpose of the LCA study, different sets of functions could be chosen. A key point shown in the table is that equating total energy use of a home with the operational phase implies a functional unit choice of all home functions. Delivering these functions includes supply chains for appliances, electronics, lights, food, and various household products, supply chains that are excluded from building LCA studies.
INTRODUCTION Life Cycle Assessment and Urban Systems. Life Cycle Assessment (LCA) is a set of methods, tools, and data designed to estimate materials flows and assess environmental impacts over the life cycle of a product or service.1,2 At the sub-building scale, LCA is used to assess building level technologies such as water heaters3 and energy systems.4 At the building level, LCA has been used to study residences2,5−7 and office buildings.8,9 Most of these studies indicate that the operational energy use overwhelmingly dominates energy use for materials and construction (e.g., 90−95% of energy use in operation versus 5−10% for materials and construction). Keoleian7 found that materials/construction share increased from 9.4% for a standard home in Michigan to 26% for an energy efficient home. At a larger scale, Norman10 compared life cycle energy of low and high population density areas in Toronto and found a substantial reduction in carbon overhead for high density living. LCA has also been used to assess urban transportation systems11 and work/lifestyle models such as telecommuting.12,13 At the level of a complete urban system, researchers have been working to assess the life cycle environmental footprints of an urban area that include impacts from producing imported goods.14 In this analysis we focus on the building level and propose methodological developments in three aspects of conducting LCA: defining the functional unit, incorporating technological progress, and parametrization. © 2011 American Chemical Society
Received: Revised: Accepted: Published: 1782
July 1, 2011 December 19, 2011 December 20, 2011 December 20, 2011 dx.doi.org/10.1021/es202202q | Environ. Sci. Technol. 2012, 46, 1782−1788
Environmental Science & Technology
Article
Table 1. Functions Associated with Residences with Associated Building Systems and Supply Chainsa function enclosed space heating/cooling lighting cleanliness food ICT and entertainment a
building systems
equipment/appliances
other products
structural structural/electrical electrical electrical/plumbing electrical/plumbing electrical
HVAC light fixtures water heater, clothes washer, dishwasher, dryer oven, microwave, refrigerator TV, computer, telephone, ....
light bulbs soap, detergent, cleaners food and beverages paper, printer ink
HVAC = heating ventilation and cooling, ICT = Information and Communication Technology.
involves carrying out a set of studies for discrete parameter values and fitting a regression model.
We thus propose that the functional unit and boundaries for reference flows be chosen in a consistent way. In this case study we consider a functional unit including enclosed space and heating/cooling. We argue that this functional unit is a reasonable choice for residential LCA, since these two functions closely connect with most building design and engage most building systems. It is not the only choice, depending on the purpose of the study, lighting, cleanliness, or other functions could be added. Note also that our scope of operational energy is narrower than prior definitions; heating cooling and ventilation (HVAC) equipment in the Unites States accounts for 52.2% of the operational energy end-use in residences.17 The share of energy used in building manufacturing relative to operation will increase with this new definition. Technological Change. Technological change presents a challenge for environmental systems assessment using LCA. LCA is usually retrospective; recommendations based on the past can be overturned by technological development. Processes evolve over time. For example, the embodied carbon in manufacturing silicon photovoltaic modules fell by 65% between 2002 and 2010.18 Products and their use also change over time. Trends in life cycle impact per functionality delivered can be dramatically different from trends in impact per typical product.19 It is thus important to attempt to account for technological progress in life cycle assessment. One approach to address dynamics is forecasting of retrospective trends in materials flows.20 Another approach is to construct future technology scenarios and then relating such scenarios to materials flows.21 Buildings are particularly interesting from the perspective of technological change. While the core of a building is generally very long-lived (e.g., 50−100 years), key elements of the technology system from an energy perspective, such as HVAC equipment, have a much shorter lifespan. Efficiency improvements in HVAC equipment have been substantial in the past decade,22 thus replacement of equipment can significantly affect life cycle energy. Parameterization of LCA. LCA is data and labor-intensive. To address this challenge there is a stream of work in the LCA community to make LCA easier to implement, such as streamlined LCA,23 scoping LCA, and development of standardized databases.24,25 Another strategy involves the use of parametric models. Within a given category of products, rather than conduct a new LCA study for each different model, a parametric model could in principle map product characteristics to life cycle impacts, i.e.
■
CASE STUDY: SCALING BEHAVIOR OF LIFE CYCLE ENERGY OF PHOENIX RESIDENCES We undertake a case study analyzing life cycle energy use of residences that incorporates our definition of the functional unit, accounts for technological progress, and estimates scaling behavior of typical residential units in Phoenix. By scaling we mean how energy used to construct and operate a residence changes as a function of its area. To control for the effects of climate and technology, we limit the study to the Phoenix metropolitan area and to a standard set of materials and technologies typically used in residential structures within this region, in compliance with local building code and the 2009 International Energy Conservation Code for Arizona’s climate zone 2. For this analysis we consider houses typical in Phoenix: average construction quality, without basement, built on a cement slab foundation, exterior walls made of stucco on a wood frame, with a cement tiled roof. Utility infrastructure and access related developments (e.g., roads, driveways) are not included in the analysis. We consider one- and two-story detached houses with area ranging from 1,500 to 3,500 square feet. We apply the following process to parametrize our LCA model: 1 Define the functional unit as delivering climate controlled lighted spaces over the assumed lifespan of the home (50 years). 2 Using the hybrid LCA method described in the next section, we estimate life cycle energy and carbon associated with materials, construction, and operation of the residence and associated equipment built in 2002. 3 Forecast technological progress in HVAC technology from 2002 to 2051 and integrate into estimation of life cycle energy.
■
4 Create a parametric model that extrapolates life cycle energy and carbon for any home in Phoenix in the 1,500−3,500 square feet range and one- or two-stories.
METHOD: COST-BREAKDOWN ECONOMIC INPUT-OUTPUT LCA LCA is a quantitative method designed to assess the environmental impacts of a product or service including relevant phases of the entire life cycle from mining, materials production, assembly, and distribution to use and disposal. LCA can be conceptually divided into inventory and assessment phases. The inventory phase includes a description of the materials and their energy use and emissions over the life cycle. This set of material and energy quantities is termed the life cycle inventory (LCI). The three main methods for estimating
parametric model: product characteristics → life cycle inventory There are continuous (e.g., area of building) and discrete (e.g., materials type) parameters that describe a product or service. The general method for constructing a parametric model 1783
dx.doi.org/10.1021/es202202q | Environ. Sci. Technol. 2012, 46, 1782−1788
Environmental Science & Technology
Article
approaches to achieve this. The simplest is additive hybrid in which economic data are identified covering processes for which materials data are unavailable and associated with sectors in an EIO model.20 An economic-balance hybrid calculates the value-added covered in a materials process model, subtracts this from the total price, and estimates impacts associated with the remaining value using EIOLCA.32 A mixed unit hybrid model constructs a matrix with both physical and economic quantities.33 We use a variant of the additive method that relies entirely on the EIO model to estimate supply chains for manufacturing. We term this method cost-breakdown EIOLCA and is based on economic analysis to account for the full cost of a product, that is, costs of different materials and basic manufacturing processes.34 Each cost element associated with an EIO sector and the total energy to manufacture a residence is given by
life cycle inventories of material and energy used are processsum, economic input-output, and their combination, known as hybrid analysis. The process-sum method, upon which most existing LCI are based, includes both a calculation method and the type of data normalization used.1 The method starts with a process network diagram, for which materials input-output data has been collected for each element in the network. The flows between processes are usually described in material terms, e.g. kg of emissions per unit mass of product output. The sources of data are often facility based, though sometimes reflect industry or even national averages. The net materials use and emissions associated with a unit output of a product or service being studied is obtained using a linear increment of materials flow associated with each process. Economic input-output life cycle assessment (EIOLCA) is based on Wassily Leontief’s formulation of an economy as a matrix describing economic transactions between sectors.26 The core of the model is the input-output matrix, usually denoted by Zmn, which describes the economic purchases and sales between economic sectors. This matrix, being a national aggregate of results of (confidential) firm-level surveys, is normally formulated by a government agency, such as the Bureau of Economic Analysis in the U.S. The most detailed tables divide an economy into 400−500 sectors. While originally formulated to address economic questions, the model can also be supplemented with environmental information to estimate supply chain materials use and emissions for products. This method has been used since the 1970s to estimate the net energy cost of products and facilities and more recently expanded to cover a broad variety of emissions and impacts.27−29 The basic formula used to calculate the net materials use or emissions associated with a unit of economic output for economic sectors is
ESC = ED(1 − A)−1
Eproduction =
∑ Ci·Esc , i i
(2)
where the subscript i refers to the ith line item in the cost model, Ci ($) refers to the cost of line item i, and Esc,i (MJ/$) refers to the corresponding supply chain energy intensity from EIO-LCA.29 The energy use in operation is estimated using the process-sum method. Technological Progress. One of the objectives of the analysis is to estimate how changes in the efficiency of HVAC equipment replaced periodically over the building lifespan would affect operational energy use. Since we consider a residence built in 2002 with a lifetime of 50 years, this estimation necessarily involves technological forecasting. There are many possible paths to approach this forecasting; here we take a scenario approach that relies on expert judgment on the future of efficiency improvements. In particular, in the Annual Energy Outlook (AEO), the Energy Information Administration of the U.S. Department of Energy report (DOE) estimates for home technology efficiencies to the year 2030.22 Forecast improvements in electric heat pumps and air conditioners are nearly the same, starting with 2.3% annual improvement in 2003, decreasing to 1.1% in 2015, then down to a 0.16% annual improvement in 2030. We use this scenario, extending it to 2052 by assuming a 0.15% annual improvement reported from 2030 until 2052. The SEER rating of the Air Conditioner stock in 2052 would be 14.4, which in our view is a pessimistic forecast. A more sophisticated treatment is a topic for future research; our purpose here is simply to scope the importance of incremental technological progress as it is currently viewed by the Department of Energy. Data and Analysis. To estimate life cycle energy and carbon associated with materials, construction, and operation of the residence and associated equipment we organized the data in the following three categories: 1 construction materials;
(1)
The result, ESC, is the vector of sector level supply chain energy use intensities (MJ/$). A is the requirements matrix built from transaction matrix (Amn = Zmn/Total economic output of sector n). ED represents direct energy use of a sector and is constructed from national (or sometimes process level) information by LCA practitioners. In the U.S., researchers at Carnegie Mellon have developed and maintained a public use model based on the 428-sector Benchmark U.S. IO tables.29 The LCA result for energy use for manufacturing the target product is found by 1 identifying its representative sector in the input-output table, 2 calculating ESC for the materials/emissions of interest, 3 multiplying by the producer price of the product (or the consumer price, depending on the formulation of the input-output model) Process-sum LCA inevitably excludes some processes in the supply chain for which materials input-output data are unavailable, leading to truncation-error. EIOLCA is a coarse grain model often combining many different processes into economic sectors, leading to aggregation error. Hybrid LCA combines the process-sum model with EIOLCA30−32 with the goal of reducing cutoff error in the former and aggregation in the latter.19 The term hybrid generically refers to any method combining process and EIOLCA; there are a number of
2 construction processes; and 3 operational data. The next section provides more detail about these components, their respective data sources, and the resulting analyses. Construction Materials. Construction material and labor costs are affected by a number of variables. These include design, building quality, area, type of materials, project location, market conditions, among other variables. For this analysis the 1784
dx.doi.org/10.1021/es202202q | Environ. Sci. Technol. 2012, 46, 1782−1788
Environmental Science & Technology
Article
business done and spent $673 million in energy purchases. Using 2002 average industrial prices and EIA’s energy conversion factors we estimated that the total energy use in this sector is equivalent to 112 PJ (petajoules), which is equal to 1.81 mega joules of primary energy per dollar of business done (see the Supporting Information). To estimate the energy used in the construction processes of a single-family house the following equation was used
economic data associated with material quantities, labor, and equipment costs for a standard building design of average construction quality is prepared using a construction estimation software. RSMeans CostWorks 2010 is used as the primary source of economic data for materials and processes associated with construction. RSMeans is a standard source of data in North America for construction cost information for both new building construction and renovation projects. The reported costs are adjusted to reflect local material costs and labor rates with the help of RSMeans’ own city Construction Cost Index (CCI) and open-shop labor rates.35 The output produced includes an up-to-date bill of material and unit costs for each material assembly of the residential structures (e.g., foundation, framing, roofing, etc.). In addition to the standard material list provided by RSMeans, we added a line item to the output to account for the manufacturing of the tools and equipment used during the construction phase. Typically, tools and equipment are not purchased and completely used-up during the construction of just a single home, but are used in multiple projects. Thus, the economic value associated with the manufacturing of the tools, based on available industry survey data of general overhead expenditures, is estimated as 1.5% of total material cost plus 5.9% of total labor cost.35 Since the EIO model is based on the 2002 Benchmark U.S. producer sector tables, we adjusted each line item in the output to reflect only the producer prices in 2002 real dollar values by multiplying each output line by a US Producer Price Index (PPI) and by a producer/purchaser ratio. We interpreted costs items such as profit, overhead, and markups as part of the wholesale portion of the purchaser price and so subtracted these to obtain the producer price. To reflect the above adjustments, eq 2 for each cost element associated with an EIO sector and the total energy to manufacture a residence is now given by Ematerial =
Econstr . process = (BVhouse / BVsector )· ∑ Fi , sector · ki i
where the subscript i refers to a specific fuel type used in construction sector, BVhouse ($) refers to the business value of the home, BVsector ($) refers to the entire business value the new single-family general contractors businesses, Fi ($) refers to total fuel purchases, and ki refers to the consumption heat rate of a given fuel type. Operational Data. The energy required for space heating and cooling was calculated using the Home Energy Saver (HES) Web-application, which uses the DOE-2 program developed by the U.S. Department of Energy for building energy analysis.37 The program provides hourly energy use and energy cost of a building given hourly weather information and a description of the building envelope and its HVAC equipment. For our analysis we customized the building envelope to match the dimensions, materials, and other prescriptive building requirement commonly found in the Phoenix area. Other general characteristics and values relating to behavior and user preferences were left in the default setting. The default settings are based on results of the Residential Energy Consumption Survey (RECS) from the Energy Information Administration.38 The RECS survey represents national household energy consumption and expenditures based on a national area-probability weighted sample of more than four thousand households. The data include detailed physical characteristics of the housing unit, appliances information, heating and cooling equipment, socio-demographics characteristics, fuel types, and quantities used. The output of the HES model is reported as on-site consumption and does not include the overhead energy input needed to produce electricity from multiple energy sources nor transmission losses. We convert all energy to primary energy to account for losses in electricity generation.39 Since the mix of primary energy used to generate electricity varies greatly across the U.S., we chose to calculate primary energy of electricity using the Western Electricity Coordinating Council (WECC) a primary energy factor per unit of delivered electricity of 2.894.40 As with prior studies we work only with average operational energy. It is important to note however that there is significant variability household by household. For example, 2005 RECS microdata from several thousand survey respondents for electricity use in a 1,500 sq. ft. home varies from 7,500 to 24,000 kWh.38
∑ (C2010, i)·(PPI2002 / PPI2010)·PPR i·Esc , i (3)
i
(4)
th
where the subscript i refers to the i line item in the cost model, and Ci ($) refers to the cost in the year 2010 from RSMeans, after subtracting installation related overhead, profits, markups, and fees in construction. PPI refers to the average US Producer Price Index for a cost item in a given year, and PPRi refers to the producer/purchaser ratio for the corresponding material producing industry that allows to estimate the cost at the material factory gate. E sc,i (MJ/$) refers to the corresponding supply chain energy intensity from EIO-LCA.29 Construction Processes. The second input to our hybrid EIOLCA model accounts for the energy utilized on-site by power tools and machinery during the construction phase. The economic value of the onsite energy use accounts for fuel consumption including gasoline, diesel fuel, and lubricants, and electric energy purchased from other companies or received from other establishments of the company. Also included are costs for natural gas, manufactured gas, fuel oil, and coal and coke products. Aggregate data on energy and resources used in the construction phase are available from the 2002 Economic Census report.36 Detailed statistics for new single-family general contractors businesses grouped in NAICS sector 236115. According to the Census data the new single-family general contractors businesses reported $62.2 billion in value of
■
RESULTS Table 2 provides a summary of the results for materials and construction for one-story dwellings of different areas, along with space heating and cooling, and operational primary consumption. We found that the total energy use in the material manufacturing and construction processes decreases, as the area of the unit increases, from 6.45GJ/m2 to 5.34 GJ/m2, for a one-story unit, and from 5.79 GJ/m2 to 4.85 GJ/m2, 1785
dx.doi.org/10.1021/es202202q | Environ. Sci. Technol. 2012, 46, 1782−1788
Environmental Science & Technology
Article
Table 2. Embedded Primary Energy in Operation, Materials, and Construction for a One-Story Residence annual operational phase
material mfg. and construction phase
area [sq.ft.]
total primary energy use [GJ]
heating and cooling [GJ]
material mfg. [GJ]
construction process [GJ]
total [GJ]
intensity [GJ/m2]
1,500 2,000 2,450 3,000 3,500
180 195 211 228 244
73.2 89.9 105 121 138
702 878 1,030 1,210 1,370
197 241 284 330 367
899 1,120 1,310 1,540 1,730
6.45 6.02 5.78 5.54 5.34
Figure 1. Embedded energy in manufacturing materials for one- and two-story 2,450 ft2 residences.
Figure 2 provides a direct comparison between three methodological variants in LCA: 1) including all primary
for a two-story unit. As dwelling area increases, the total life cycle embedded energy in the construction and material processes grows nearly linearly. These results are consistent with the observation that material requirements, such as cement used for the foundations, scale proportionally with the overall area of the building, while other materials such as drywall, paint, wood, and stucco scale as a function of the perimeter or volume. Figure 1 shows embedded energy in materials used in oneand two-story residences of the same area: 2,450 square feet. Note that 10−15% energy savings can be obtained when building a two- rather than a one-story unit. The smaller energy footprint of the two-story house is because lower energy materials such as wood substitute for more energy intensive materials such as cement. Materials included in the “others” category contribute far less individually to the total embedded energy and did not vary significantly in quantities for single- and two-storied dwellings. With our definition of functional unit and after accounting for technological progress we find that between 19% and 30% of the total life cycle energy can be attributed to materials and construction processes, as opposed to the 0.4−11% found in other studies.5,10,41−43 Note that our results for total energy to manufacture a 2,450 square foot residence (1,310 GJ) are similar to those found in prior studies (1,435 GJ for a 2,450 square feet standard home)7,16 Thus the difference in share is explained by the functional unit choice and not difference in method.
Figure 2. Comparison between definitions of functional unit for a onestory 2,450 ft2 residence.
energy end-use in the functional unit (with no technological improvements); 2) including only HVAC primary energy uses (with no technological improvements); and 3) our proposed functional unit. Using our definition of functional unit and our methodology, the embedded energy in construction and 1786
dx.doi.org/10.1021/es202202q | Environ. Sci. Technol. 2012, 46, 1782−1788
Environmental Science & Technology
Article
energy is associated with residences and products used in residences, it is only natural that manufacturing should represent a reasonable share of life cycle energy. The reinterpretation of functional unit has implications for planning sustainable, energy-efficient communities. In particular there is a stronger case to manage the embedded energy in building materials in building codes. Such codes will go a long way in reducing total energy demands in a community. In addition, subdivision regulations can take advantage of specific volumetric considerations that incentivize builders to build up rather than horizontally, once they hit a particular building footprint threshold. Future research needs to examine the scaling behavior of high-rise buildings with respect to energy given that there are expected discontinuities in such relationships. While this article showed only a simple parametric model mapping one product characteristic − house area − to life cycle energy, the approach is general and potentially of great utility in LCA. While LCA has been mainly approached from a case study perspective, new software-based models are being developed to allow users input individual data. For example, for buildings the e-Quest and DOE-2 models generate operational energy use based on a variety of user input data on building, equipment, and usage .45 One can envision a suite of LCA models for different products that allow the user to customize product supply chain, design, and operation characteristics.
material manufacturing processes accounts for 27.3% of the total embodied energy for a 2,450 square feet one-story dwelling. Specifically, the embedded energy used during construction corresponds to 5.7% and in the material manufacturing 21.6% of the total life cycle energy. This result is in contrast to the more typical 11.4% when all primary energy end-use are included in the functional unit, without accounting for technological improvement. Lastly, we build a parametric model mapping house area to life cycle energy. This is done by regressing the five areas studied for different functional forms on estimated total lifecycle energy. While higher order polynomials give a marginally better fit, a linear regression model was sufficient to reproduce energy for the house areas examined in our study. We found that the fit of our linear regression models for both single-story and two-story units were extremely good with R-squares in the high 90s. This result suggests that the embedded energy in residential units scale linearly with area of the unit after controlling for the number of stories. The coefficients of the area parameter suggests that for each additional square foot increase in area of livable space, for a one-story unit, the embedded energy goes up by 419 MJ. For two-story residential units this marginal increase per square foot is 360 MJ. Figure 3 shows that the linear scaling behavior across different unit areas for total embedded energy is also reflected in the disaggregated components of the total as well.
■
ASSOCIATED CONTENT
* Supporting Information S
Tables S-1−S-6. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected].
■
ACKNOWLEDGMENTS This research was supported by the Civil Infrastructure Systems program at the National Science Foundation (CMMI grant # 1031690). The authors thank Ariane Middel for helpful input.
■
Figure 3. Life cycle primary energy embedded in materials, construction, and HVAC for one-story residence of different areas, with linear parametrizations.
■
REFERENCES
(1) Baumann, H.; Tillman, A.-M. The Hitch Hikers Guide to LCA; Studentlitteratur AB: Lund, Sweden, 2004. (2) Hendrickson, C. T.; Lave, L. B.; Matthews, H. S. Environmental Life Cycle Assessment of Goods and Services: An Input-Output Approach; Resources for the Future Press: 2006. (3) Taborianski, V.; Racine, P. Comparative evaluation of the contribution of residential water heating systems to the variation of greenhouse gases stock in the atmosphere. Build. Environ. 2004, 39 (4), 645−652. (4) Osman, A.; Ries, R. Life-cycle impact analysis of energy systems for buildings. Infrastruct. Syst. 2005, 10 (3), 87−97. (5) Adalberth, K. Energy use during the life cycle of single-unit dwellings: examples. Build. Environ. 1997, 32 (4), 321−329. (6) Fay, G. R.; Treloar, G.; Iyer-Raniga, U. Life-cycle energy analysis of buildings: a case study. Build. Res. Inf. 2000, 28 (1), 31−41. (7) Keolian, G.; Blanchard, P. Reppe. Life cycle energy, costs and strategies for improving a single family house. Ind. Ecol. 2000, 4 (2), 135−157. (8) Cole, R.; Kernan, P. Life-cycle energy use in office buildings. Build. Environ. 1996, 31 (4), pp307−317.
DISCUSSION
This study finds that by designating a different functional unit, the contribution of materials and construction to the life cycle energy of a residence is far higher than previous studies. Prior studies excluded potentially important supply chains (e.g., appliances, food, household chemicals) associated with the implicit functional unit. We argue that the conventional wisdom that operational energy use overwhelmingly dominates the life cycle of buildings needs to be reconsidered. This higher share is also more intuitive from a macroperspective. Since industry represents 20% of U.S. energy demand,44 and much of this 1787
dx.doi.org/10.1021/es202202q | Environ. Sci. Technol. 2012, 46, 1782−1788
Environmental Science & Technology
Article
assessment and material flow analysis. Environ. Sci. Technol. 2007, 41 (3), 1024−1031. (34) Bullard, C.; Pennter, P.; Pilati, D. net energy analysis: handbook for combining process and input-output analysis. Resour. Energy 1978, 1, 267−313. (35) RSMeans Building Construction Cost Data 2010; RSMeans Company Inc.: Kingston, MA. (36) 2002 Economic Census Industry summary: Construction; Subject Series ECO02-23SG-1. Economic and Statistics Administration: Washington, DC. http://www.census.gov/prod/ec02/ec0223sg1.pdf. (37) Lawrence Berkeley National Laboratory. The Home Energy Saver (HES). http://hes.lbl.gov (accessed February 2011). (38) Residential Energy Consumption Survey 2005; U.S. Energy Information Administration (EIA). http://www.eia.gov/consumption/ residential/data/2005/ (accessed June 2010). (39) DOE: 2007 Buildings Energy Data Book; U.S. Department of Energy: Washington, DC. http://buildingsdatabook.eren.doe.gov/ docs%5CDataBooks%5C2007_BEDB.pdf. (40) Deru, M.; Torcellini, P. Source energy and emission factors for energy use in buildings. Technical Report; NREL/TP-550-38617; 2007. (41) Hellwig, R.; Erhorn, H. Life cycle assessment of a prefabricated timber-frame house. International Symposium of CIB W67, Energy and Mass Flow in the Life Cycle of Buildings; Vienna, 1996; pp 321−327, 4−10. (42) Lippke, B.; Perez-Garcia, W. Bowyer and Meil, CORRIM: lifecycle environmental performance of renewable building materials. For. Prod. 2004, 54 (6), 8−19. (43) Guggemos, A.; Horvath, A. Comparison of environmental effects of steel- and concrete-framed buildings. Infrastruct. Syst. 2005, 11 (2), 93−101. (44) Energy Consumption by Sector and Source, United States, Reference Case; U.S. Energy Information Administration (EIA): 2011. http:// www.eia.gov/oiaf/aeo/tablebrowser/ (accessed June 2010). (45) Hirsch, J. and Associates. DOE-2.2 - Building Energy Use and Cost Analysis Program Vol. 6: New Features, Version 41−47; Camarillo, CA, 2009.
(9) Junnila, S.; Horvath, A. Life-cycle environmental effects of an office building. Infrastruct. Syst. 2003, 9 (4), 157−163. (10) Norman, J.; MacLean, L.; Kennedy, C. Comparing high and low residential density: life-cycle analysis of energy use and greenhouse gas emissions. Urban Plann. Dev. 2006, 132 (1), 10−21. (11) Chester, M.; Horvath, A. Environmental assessment of passenger transportation should include infrastructure and supply chains environmental. Environ. Res. Lett. 2009, 4, 1−4. (12) Kitou, E.; Horvath, A. Energy-related emissions from telework. Environ. Sci. Technol. 2003, 37 (16), 3467-3475. (13) Matthews, S.; Williams, E. Telework adoption and energy use in building and transport sectors in the us and Japan. Infrastruct. Syst. 2005, 11 (1), 21−30. (14) Ramaswami, A.; Hillman, T.; Janson, B.; Reiner, M.; Thomas, G. A demand-centered, hybrid life-cycle methodology for city-scale greenhouse gas inventories. Environ. Sci. Technol. 2008, 42 (17), 6455−6461. (15) Mithraratne, N.; Vale, B. Life cycle analysis model for New Zealand houses. Build. Environ. 2004, 39, 483−492. (16) Ochoa, F.; Ries, R.; Matthews, H. S. Hendrickson, C. Life cycle assessment of residential buildings. In American Society of Civil Engineers Construction Research Congress, San Diego, CA, 2005. (17) DOE 2009 Buildings Energy Data Book; Residential Sector Energy Consumption Table 2.1.5; U.S. Department of Energy: Washington, DC. http://buildingsdatabook.eren.doe.gov/ChapterView. aspx?chap=2 (accessed January 2011). (18) Zhai, P.; Williams, E. Dynamic hybrid life cycle assessment of energy and carbon of multi-crystalline silicon photovoltaic (PV) systems. Environ. Sci. Technol. 2010, 44 (20), 7950−7955. (19) Deng, L.; Williams, E. Functionality versus “Typical Product” measures of technological progress: case study of semiconductor manufacturing. Ind. Ecol. 2011, DOI: 10.1111/j.1530-9290.2010.00306.x. (20) Williams, E.; Weber, C.; Hawkins, T. Hybrid approach to managing uncertainty in life cycle inventories. Ind. Ecol. 2009, 15 (6), 928−944. (21) Pehnt, M. Dynamic life cycle assessment (LCA) of renewable energy technologies. Renewable Energy 2006, 31, 55−71. (22) Supplemental Tables to the Annual Energy Outlook 2008. DOE/ EIA-0383, Table 21; Energy Information Administration, U.S. Department of Energy: Washington, DC. http://www.eia.gov/oiaf/ archive/aeo08/supplement/index.html. (23) Graedel, T. Streamlined Life Cycle Assessment; Prentice Hall: New York, 1998. (24) Swiss Center for Life Cycle Inventories. EcoInvent database, version 2.2; St. Gallen, Switzerland, 2011. (25) Athena Institute. ATHENA Impact Estimator for Buildings. Version 4.1; Merrickville, Ontario, Canada, 2011. (26) Leontief, W. Environmental repurcussions and economic structure - input-output approach. Rev. Econ. Stat. 1970, 52 (3), 262−271. (27) Bullard, C.; Herendeen, R. The energy cost of goods and services. Energy Policy 1975, 268. (28) Hendrickson, T.; Horvath, A.; Joshi, S.; Lave, L. B. Economic input-output models for environmental life-cycle assessment. Environ. Sci. Technol. 1998, 32 (4), 184A. (29) Green Design Institute, Carnegie Mellon University. Economic input-output life cycle assessment. 2011. http://www.eiolca.net (accessed January 2011). (30) Engelenburg, W.; Van Rossum, M.; Blok, K.; Vringer, K. Calculating the energy requirements of household purchases: a practical step by step method. Energy Policy 1994, 21 (8), 648−656. (31) Suh, S.; Lenzen, M.; Treloar, G.; Hondo, H.; Horvath, A.; et al. System Boundary Selection in Life-Cycle Inventories Using Hybrid Approaches. Environ. Sci. Technol. 2004, 38 (3), 657−664. (32) Williams, E. Energy intensity of computer manufacturing: hybrid assessment combining process and economic input-output methods. Environ. Sci. Technol. 2004, 38 (22), 6166−6174. (33) Hawkins, T.; Hendrickson, C.; Higgins, C.; Matthews, H. S.; Suh., S. A mixed-unit input-output model for environmental life-cycle 1788
dx.doi.org/10.1021/es202202q | Environ. Sci. Technol. 2012, 46, 1782−1788
