review the thesis

Sareza1027

11.pdf

Home >Architecture and Design homework help >review the thesis

R E S E A R C H A N D A N A LY S I S

Construction Matters Comparing Environmental Impacts of Building Modular and Conventional Homes in the United States

John Quale, Matthew J. Eckelman, Kyle W. Williams, Greg Sloditskie, and Julie B. Zimmerman

Keywords:

environmental impact assessment industrial ecology life cycle assessment (LCA) modular building off-site construction residential homes

Summary

Modular construction practices are used in many countries as an alternative to conventional on-site construction for residential homes. While modular home construction has certain advantages in terms of material and time efficiency, it requires a different infrastructure than conventional home construction, and the overall environmental trade-offs between the two methods have been unclear. This study uses life cycle assessment to quantify the environ- mental impacts of constructing a typical residential home using the two methods, based on data from several modular construction companies and conventional homebuilders. The study includes impacts from material production and transport, off-site and on-site energy use, worker transport, and waste management. For all categories considered, the average impacts of building the home are less for modular construction than for conventional con- struction, although these averages obscure significant variation among the individual projects and companies.

Introduction

In the United States, buildings represent the largest single end-use of energy and emitter of greenhouse gases (GHGs). Quantitatively, buildings account for 40% of energy use in the United States and a comparatively significant proportion else- where (Pérez-Lombard et al. 2008; DOE 2008). Energy and materials are used, and corresponding environmental impacts incurred, in large quantities throughout the life cycle of a build- ing. There has been significant life cycle assessment (LCA) research dedicated to identifying which materials and building components are the largest contributors to various environ- mental impacts (Ortiz et al. 2009); however, the process of

Address correspondence to: Matthew J. Eckelman, Department of Civil & Environmental Engineering, Northeastern University, 360 Huntington Ave., Boston, MA, USA 02115. Email: [email protected]

Volume 16, Number 2

conducting an LCA on buildings is complicated. Factors such as site specifications, thousands of potential components and material types, and various construction assembly techniques (Priemus 2005), as well as the highly decentralized nature of the industry, make it very difficult to track down the neces- sary data and produce meaningful, generalizable results (Kohler and Moffatt 2003; Malin 2005; Sharrard 2007). Nevertheless, lessons learned from various LCAs are broadly applicable and it is clear that LCA is a useful and even necessary component of a holistic and integrated green building design process (Ortiz et al. 2009).

It is well established that the largest proportion of envi- ronmental impacts associated with buildings is related to the

www.wileyonlinelibrary.com/journal/jie Journal of Industrial Ecology 243

R E S E A R C H A N D A N A LY S I S

occupancy or use stage of the life cycle (Adalberth 1997; Scheuer et al. 2003). For example, energy used for heating, cooling, lighting, equipment, and appliances typically far out- weighs the energy demand of other life cycle stages, such as construction and the production of building materials. These proportions will likely change with time as building designs and operations improve in terms of energy and material efficiencies. Full implementation of current design strategies and technolo- gies could see the proportion of impacts associated with the occupancy phase fall sharply. As this happens, the relative im- portance of other life cycle stages such as construction will be greater. This point is illustrated by a recent study of Gustavs- son and Joelsson (2010), who show that in an optimally energy efficient building, the phases of material production and con- struction account for 60% of life cycle energy consumption. As such, it is necessary also to begin developing and evaluating strategies to reduce the impacts associated with these stages of the building life cycle.

Modular building is one promising technique to lower the impacts of construction and is utilized for various building types, including single-family homes, multifamily housing, hotels, dor- mitories, and various commercial and retail structures. Modu- lar construction is a form of prefabrication that involves the creation of discrete volumetric sections of buildings that are transported to a site and assembled into a complete building. The modules are produced off-site in a factory environment without exposure to weather. Panel systems are a related off- site construction technique, where individual wall sections are manufactured without factory assembly into volumetric mod- ules. Unlike in panelized or component-based methods of pre- fabrication, in modular construction most of the interior and exterior finishes are put into place in the factory. The modules are transported to the building site 80% to 90% complete, where they are assembled and finished. Because the modules can be constructed while the site and foundation are being prepared, instead of after, modular construction is thought to reduce con- struction times by 30% to 50% (Smith 2010). Modular compa- nies may also economize on costs by producing several modules in parallel, reducing worker and machinery transportation, and engaging in bulk material ordering.

Residential modular buildings are different than trailers or “double-wide” homes, which are defined as “manufactured hous- ing” by the federal government. Trailers have a permanent steel frame built into the floor structure, and can be relocated. Con- sidered temporary buildings because they are transportable, they are built to federal building requirements known as the HUD (U.S. Department of Housing and Urban Development) code, and are financed like a vehicle. Unlike site-built or modular homes, they also tend to depreciate in value. However, mod- ular houses are built to local and state building codes, in the same way as a site-built house, and those codes are typically more restrictive than the HUD code on energy and durability issues. A modular house is of equal quality to a site-built home because the materials are exactly the same for the same design, with the exception of added structure to ensure the modular house can be transported without being damaged.

The size of the modular industry nationally in the United States is approximately 2% to 3% of residential construction, with wide variation among regions. In the Northeast, nearly 6% of new homes utilize modular techniques, while the pro- portion in the West is less than 1%. The broader category of prefabricated construction (modular, manufactured, and panel- ized homes) represents more than 25% of the total (Hallahan Associates 2011).

Some aspects of modular home construction are identical to conventional practices, such as site preparation, excavation, and installation of the foundation, but there are many impor- tant differences. While building homes in a factory can require additional lighting, heating, or water use in addition to the en- ergy associated with transporting the modules, there are also opportunities for benefits in terms of material and energy effi- ciencies. As such, the environmental trade-offs between mod- ular building and conventional on-site construction have not been thoroughly investigated.

Accordingly, the aim of this study is to compare these two methods of building construction using LCA. The analysis is based on data from several modular and on-site building projects in the mid-Atlantic United States and focuses on the dif- ferences between the two construction techniques, covering production and transport of building materials, construction processes, worker transport to the job site and/or factory, and construction waste management. Although panelized construc- tion is more popular in some parts of the United States and Europe, this report focuses on modular prefabrication. This is a cradle-to-gate study that considers only the materials pro- duction, transport, and construction phases of the building life cycle. There are potential differences in the use phase between homes constructed using modular and conventional techniques, due to envelope tightness, for example, but such effects are highly variable and were not included in this construction- focused study.

Life Cycle Assessment Research on Buildings and Construction

Researchers have used LCAs to examine many aspects of buildings and construction. The majority of these studies have analyzed energy consumption and, to a lesser extent, global warming (Ortiz et al. 2009), whereas very few have addressed other environmental concerns such as loss of habitat or emis- sions of toxic substances. One line of inquiry has focused on the impacts of material or design choices in the life cycle of build- ings using process-based LCA. An early comparison of wood, steel, and concrete office buildings conducted by Cole and Ker- nan (1996) quantified life cycle energy use and found little difference in impact between the different material structural systems, especially on the scale of the entire life cycle. Keoleian and colleagues (2000) considered GHG emissions as well in a study of a residential home in Michigan, analyzing the life cycle gains and losses from several green building options. Itard and Klunder (2007) compared the life cycle embodied materials, energy, and water use among four different options of building

244 Journal of Industrial Ecology

R E S E A R C H A N D A N A LY S I S

renovation ranging from simple maintenance to complete demolition and rebuilding. They found that although demo- lition and rebuilding have large impacts in terms of embodied materials, the lower use phase energy and water impacts of a new building mean that there are large trade-offs to be considered. Another collection of LCA building studies has used economic input–output life cycle assessment (EIO-LCA) or hybrid anal- yses to model economy-wide life cycle impacts of a building, or a subset of building materials and components, again typically considering energy and emissions of GHGs (Bilec et al. 2006; Ochoa et al. 2002; Treolar et al. 2000).

A number of building construction LCA studies have ana- lyzed environmental impacts other than energy use and GHG emissions. Scheuer and colleagues (2003) completed one of the most comprehensive process-based LCAs for an institutional building, finding that impacts of global warming, ozone deple- tion, eutrophication, and acidification potentials followed the same basic pattern as energy use where the use phase of the building was the dominant source. The only impact category that approached parity between the use phase and other life cy- cle stages was waste production, with material production and construction accounting for approximately 28%, as compared to 65% generated during the use phase (Scheuer et al. 2003). Guggemos and Horvath (2005) found that the construction phase accounted for only 2% of energy consumption and 1% of GHG emissions, but 7% of carbon monoxide (CO) emissions, 8% of nitrogen oxides (NOx) emissions, and 8% of particulate matter (PM) emissions. In some cases, the construction phase was found to contribute the majority of impacts for certain impact categories. For example, Ochoa and colleagues (2002) found that the construction phase contributes 57% of toxic air emissions and 51% of hazardous waste generated over the build- ing life cycle. Similarly, Junnila and Horvath (2003) found that the impacts of producing the building materials dominated the release of toxic metals to the environment and also contributed significantly to smog-forming emissions.

In summary, the literature to date suggests that the use or occupancy phase tends to dominate most environmental im- pacts over the life cycle of a building, especially for energy use and GHG emissions, with some exceptions for other impact categories. However, another significant conclusion is that the construction phase is the next most important area in terms of impacts, and therefore the potential for reducing the environ- mental burden of buildings. For example, while the occupancy phase is reported to account for anywhere from 70% to 98% of building energy use (Ortiz et al. 2009), the construction phase has been found to account for 2% to 26%, depending on the reference building’s design and intended use (Adalberth 1997; Cole 1998; Cole and Kernan 1996; Keoleian et al. 2000; Oregon Department of Environmental Quality 2010; Scheuer et al. 2003).

There are many potential sources of construction phase im- pacts that are discussed. Gangolells and colleagues (2009) found that transportation and construction equipment, waste produc- tion, and water consumption all had significant environmental consequences, implying that any improvements in these areas

could be priority targets for reducing the overall life cycle im- pact of the building. Bilec and colleagues (2006) similarly found for a concrete parking structure that the transportation of con- crete to the site, particularly for the precast pieces, was the most significant construction process. Likewise, Guggemos and Hor- vath (2005) found that within the construction phase, work on the structural frame had the largest impacts, mostly due to the heavy use of diesel equipment, a finding that was corroborated by Junnila and Horvath (2003).

Very few researchers have attempted to analyze the impact of the method of construction. In particular, there is little dis- cussion of the implications of off-site construction as opposed to typical site-built construction. Kim (2008) compared life cycle impacts of a modular and a conventionally constructed home in Michigan, analyzing energy use, material consumption, GHG emissions, and waste generation. This work suggested that solid waste generation, transportation energy, and GHG emis- sions are significantly lower when modular construction is used. For example, solid waste generation was found to be 2.5 times greater for the on-site construction process. However, this study focused on a single building and relied on several general as- sumptions to fill gaps in empirical data. Another relevant study was carried out in northern Japan by Nishioka and colleagues (2000), who considered the environmental trade-offs in ver- tically integrated factory-built housing, which required more materials than a typical home but also performed better. The authors found that the energy and carbon debts incurred by the additional materials were paid off through efficiency gains in less than six years, well below the average lifetime of homes in that area.

With increasing interest in the environmental profile of con- struction and green buildings generally, the lack of quantitative assessment work on modular construction is an important omis- sion in the existing literature. In addition to environmental considerations, stakeholders from industry and government are also interested in the relationship between modular construc- tion and the affordability and availability of homes (Britto et al. 2007; Diez et al. 2007).

Methodology

The first step of the project was to collect construction data for both off-site and on-site construction methods. Three resi- dential modular companies, generally representative of the east- ern U.S. modular industry, supplied data on completed projects for this study, including utility bills, worker commuting infor- mation, building materials and waste procedures, construction schedules, employee schedules, and other relevant information. As most data were reported as an amount per week or year, an- nual production estimates for each modular building factory were used to scale the information to the common functional unit of a 2,000 square foot (sq ft, or 186 square meters [m2]),1

two-story home that is a model for one of the companies in- volved in the study (figure 1).

As noted in previous studies (Kim 2008), it is difficult to identify instances where two versions of the exact same

Quale et al., Construction Matters: LC A of Modular Buildings 245

R E S E A R C H A N D A N A LY S I S

Figure 1 Elevations for the functional unit of a 2,000 square foot, two-story production model.

building have been completed using on-site and modular con- struction, making it unfeasible to make comparisons based on actual projects. Therefore it is necessary to use proxy data for one or the other method of construction in order to compare the environmental impacts of building a common functional unit. Other authors have used construction cost estimators for on-site construction processes (Bilec 2006; Kim 2008), which do not directly yield environmentally relevant information, but can be used in combination with EIO-LCA.

Here we estimate material and energy use data by compiling thorough construction specifications for the modular building shown in figure 1 and then surveying a random sample of five experienced on-site professional homebuilders in the sales re- gion of the modular building companies. The homebuilders were asked to assume they were building the home on-site at a lot in their region. They were selected because they had recent experience constructing buildings of the same size and type. The survey covered the general steps required to construct the building, from site preparation through final cleanup. These pro- fessionals were provided complete drawings and specifications of the modular building to estimate the construction schedule, equipment and energy needs, and the required number of staff and subcontractors, including their commuting information. The survey results provided the necessary data for the analysis of on-site construction.

For modular construction, a two-story modular residence can be visualized as consisting of four boxes: A, B, C, and D. Boxes A and B sit side by side and make up the ground floor, while C and D sit side by side on top of A and B and form the second floor. Due to restrictions on shipping dimensions on public roads, it is not possible to build the house depicted in figure 1 with any fewer than four separate modular pieces. Each of the four pieces that make up this house must have a top, bottom, and four sides in order to ensure structural integrity during shipment and installation. This means, however, that in a modular house there is the possibility of a number of redundant walls and floors. For the modular house design used in this

fold up roof

module A module B

module C module D

foundation

structure required for both modular and conventional construction

structure required for modular construction only; this structure can be temporary

Figure 2 Schematic diagram of a four-module residential building, showing redundant structure. The sections below detail the assumptions and parameters of each building life cycle phase considered in the study: materials production, materials transport, worker transport, construction processes (in factory and on-site), and waste management. A detailed list of material and energy inputs for each mode of construction is given in table 1.

study, there are two sets of wall studs between boxes A and B (and between boxes C and D), creating two redundant vertical surfaces. There is also both a ceiling for box A and a separate floor for box C (and between boxes B and D) resulting in two redundant horizontal surfaces (figure 2). The added material for the redundant structure was not present for the site-built version of the home in this analysis.

Materials Production

In comparing the modular and on-site methods, only those building materials whose amounts differed between construc- tion methods were considered. For example, dimensional lum- ber for wall framing was included, while doors and windows, which appear identically in both versions of the building, were not. Material types and quantities came from both bills of ma- terials and estimates based on the construction drawings for

246 Journal of Industrial Ecology

R E S E A R C H A N D A N A LY S I S

Table 1 Inputs of materials and energy for modular and site-built versions of the home functional unit (only those inputs that differ appreciably between construction methods are shown)

Construction method

Unit Modular Conventional

Material Wood (marriage walls) lb 3,190 0 Wood wastage lb 0 3,300 Drywall wastage lb 1,380–1,600 2,200

Material transport Building materials: 16 metric ton truck miles 110–480 110–1,600 Building materials: 28 metric ton truck miles 30–290 110–290 Modules to site: 28 metric ton truck miles 1,200 0

Worker transport To factory: car/light duty truck miles 2,500–13,000 0 To site: car/light duty truck miles 1,000–3,500 7,800–26,000

Energy use Electricity off-site MWh 2.0–7.2 0 Electricity on-site MWh 0.1 0.5–1.6 Gasoline (equipment) MMBTU 24 15–54 Fuel oil (heating off-site) MMBTU 24 0 Propane (heating off-site) MMBTU 0.4 0 Natural gas (heating off-site) MMBTU 0.7 0 Natural gas (heating on-site) MMBTU 10–18 85–145

Waste management Mixed materials lb 4,580 5,500

Notes: lb = pounds; MWh = megawatt hours; MMBTU = million British thermal units (Btu).

the building (figure 1). Based on surveys of contractors, it was determined that the foundation and roof structure would not appreciably differ between the two construction methods.

There are two important differences in the material require- ments of modular and on-site construction projects that were incorporated in this analysis. First is the amount of material that gets incorporated into the structure. As noted above, the marriage walls required for transporting and then joining the modules add nearly 25% to the mass of wood in the building. The second major difference in material requirements between the two construction methods is related to waste, as this rep- resents extra building materials that must be ordered to com- plete the home functional unit. In general, modular residential companies build many homes at once and order materials and products as needed for the efficient use of time and the factory’s material storage space. Many modular building companies also purchase dimensional lumber and other materials cut to spec- ifications, or make cuts using digital fabrication equipment, resulting in fewer off-cuts, and therefore requiring less over- ordering to make up for wastage. Small pieces can be saved for future projects, such as wood used for blocking. The general policy not to begin construction on a modular home until all materials have been procured also reduces over-ordering and impacts associated with multiple deliveries.

In contrast, on-site homebuilders typically order materials and products for one building at a time (for the purposes of this study the surveyed homebuilders assumed the lot was discrete

and not part of a larger housing development), and may make many ad hoc trips to building supply stores to procure materials as needed. Conventional homebuilders generally lack sufficient dry or climate-controlled storage space at the construction site to store building materials and have fewer staff to determine efficient procurement strategies. The homebuilders surveyed for this study agreed that an “order as you go” strategy is less efficient and can also lead to additional material and employee transportation impacts. There is anecdotal evidence that on- site builders routinely order a surplus of 5% to 15% to make up for wasted material. Kim (2008) assumed a wastage rate of 5% across all material types, but it is more likely that the rate varies significantly across materials. This rate is quite difficult to track, however, with subcontractors managing some waste streams (such as plumbers keeping copper pipe cuts) out of the purview of the general contractor.

A survey of modular construction facilities revealed that practically all building materials were reused, with the excep- tion of some gypsum (0.7–0.8 pounds per square foot [lb/sq ft]) and copper wire (0.03–0.1 lb/sq ft), which are imprac- tical to use in small sections.2 These amounts were derived from factory reports of waste generation, which were supplied as a weekly or annual tally and then scaled to the 2,000 sq ft functional unit. On-site construction over-ordering was as- sumed to be equal to the generic wastage rate from residen- tial construction of 4.4 lb/sq ft (EPA 2009). This was broken down into constituent material types, specifically dimensional

Quale et al., Construction Matters: LC A of Modular Buildings 247

R E S E A R C H A N D A N A LY S I S

lumber (1.7 lb/sq ft) and gypsum wallboard (1.1 lb/sq ft), based on the proportions reported in characterization studies of con- struction waste (DSM Environmental Services 2008; NAHM 1997).

The top section of table 1 shows the net differences in material requirements for the two construction methods, cov- ering both structural redundancy and material efficiency, in- cluding dimensional lumber, drywall, and masonry. Those materials types that are not expected to significantly differ in quantity between the two construction methods, including sid- ing, roofing materials, metal piping and wiring, carpets, fixtures, tiles, packaging cardboard, and others, are not considered in the comparison.

Material Transport

For conventional on-site construction, transportation im- pacts of only those additional materials needed beyond those required for modular construction were considered (table 1). The remaining majority of building materials were excluded, as the impacts of transporting these from distributors to the site was assumed to be the same as transporting them from distrib- utors to the modular building factories. An average distance of 30 miles3 from material distributors to the construction site and from the site to the local construction and demolition (C&D) waste management facility was assumed for all materials, which is similar to transport distances reported in the work of Kim (2008). The transport of waste from modular and conventional construction sites was also considered. The vehicle used for all material transportation was modeled as an average large cargo truck with empty backhauling.

For modular construction, materials are transported to the factory and then the modules themselves need to be transported to the building site. Transport requirements for the finished modules were calculated using the weight of the modules (four modules each weighing approximately 10 metric tons (t)4 for a 2,000 sq ft home) and the average shipping distance for the modular companies who participated in the study, which was estimated to be 300 mi.

Worker Transport

For on-site construction, worker transport was calculated based on data supplied in the homebuilder survey (see table 1). For each task (interior painting, for example), each respondent was asked to estimate the number of people needed, the number of days to finish the task, and city of origin of those workers. Round-trip commuting distances were calculated as

Off-site worker-miles

= Round-trip distance × People needed for task × Duration of task in days × 1.5.

The 50% factor of safety was determined from contractor interviews, as surveys indicated that scheduling delays or tasks that must be repeated are commonplace during on-site con- struction projects. For example, the plumbing subcontractor

may arrive prior to completion by the framing subcontractors or the inspector must make multiple trips to permit various as- pects of the project. This parameter is considered later in the uncertainty analysis. All on-site worker miles were modeled as small-truck transport, except in the cases where large equip- ment or material was being delivered (such as an excavator for site preparation or a cement truck). In these cases, larger truck sizes were assigned as appropriate and the trucks were assumed to be loaded to 50% of capacity for those miles.

In the case of modular construction, worker transport to the facility was determined using data on where employees live in relation to the factory and the number of workdays per year. Transportation and craning of the modules were also assessed. With the actual commuting distance of each employee, weekly days of work, and an assumption of 50 working weeks a year, total annual worker miles were calculated as

Off-site worker-miles

= Round-trip distance × Number of employees ×Workdays/week × 50 weeks.

Using annual production of a factory in total square feet of built area, a normalized worker transport metric was determined for each company, which was then scaled to the functional unit of a 2,000 sq ft building. It was assumed that miles driven by workers at modular factories were 50% by small car and 50% by van or truck, based on conversations with the modular companies and photographs of their parking lots.

Energy Use During Construction

Data on the energy used during on-site construction also came from the homebuilder surveys (see table 1). For each task, the respondent was asked to give the electric- or diesel-powered equipment that would typically be used during the construction task and the duration of use. Average fuel or electricity use for each appliance or piece of machinery came from the U.S. federal government’s Energy Star program (EPA 2011) and technical specifications from manufacturers. Energy use calculations com- bined the per hour energy usage with assumed usage per day and days for the task as

Energy use

= Hourly energy use × Hours per day used × Days per task.

Energy used by vehicles and generators was recorded as diesel in machinery with a gross heating value of 45 megajoules per kilogram (MJ/kg).5 Electricity used by tools and other electri- cal equipment was recorded as kilowatt-hours (kWh) of low- voltage electricity in the eastern connect of the U.S. grid, in- cluding transformation and distribution losses and using gross primary energy heating values, from the U.S. Life Cycle Inven- tory database (NREL 2011).

A number of assumptions were made regarding how often equipment was used to ensure a consistent methodology across the different homebuilders. For office equipment used in the general contractor’s office, daily use was assumed to be only

248 Journal of Industrial Ecology

R E S E A R C H A N D A N A LY S I S

Figure 3 Global warming potential from modular (Mod) and conventional (Conv) construction cases of a 2,000 square foot residential home, in metric tons of carbon dioxide equivalents (CO2-eq); differences in construction only.

eight hours per day, although many pieces of equipment, such as computers and printers, were reportedly left on overnight. It was also assumed that the general contractor had an office for both on-site and off-site construction cases, either as a mobile unit or as dedicated office space in the modular building factory. For on- site electricity use, it was assumed that a temporary power pole was used (i.e., grid electricity) unless a stand-alone generator was specified in the construction documents (for a concrete vibrator or an air compressor, for example). If temporary lighting was specified, it was assumed that one 100 watt (W) light per worker would be used.

One source of energy use that was not explicitly covered in the survey for the on-site projects is temporary heat on the job site. Based on the experience of the research team on con- struction sites and in further consultation with contractors, it was assumed that in the mid-Atlantic region (where the on-site contractors are based), 50% of an on-site construction schedule requires temporary heat. Of that 50% of the schedule requir- ing temporary heat, 30% requires continuous heat (to protect interior finishes, for example).

Energy use in the modular factories was determined based on energy bills provided by the companies. Annual fuel and electricity use were calculated using average yearly energy ex- penditures, scaled to the 2,000 sq ft building functional unit.

Waste Management

Modeling of waste management also considered, for com- parative purposes only, the net differences in waste between the two construction methods from table 1. All materials were assumed to be sent to a C&D waste landfill.

Life Cycle Assessment Software

The life cycle inventory was compiled in SimaPro 7.3 LCA software, relying on inventory data from the US-EI version of the ecoinvent 2.2 and U.S. LCI databases, with complete adaptation of all generic electricity unit processes to the U.S.

context. Data from the five categories described above were entered as separate items for each of the three modular and five on-site cases. Subsequent impact assessment was performed using the Building for Environmental and Economic Sustain- ability (BEES) 4.02 method, which was developed specifically for the U.S. building sector (NIST 2007). This yielded impact results for ten environmental impact categories: global warm- ing; acidification; human health from cancer, non-cancer, and criteria air pollutants; eutrophication; ecotoxicity; smog; water intake; and ozone depletion.

Uncertainty

Several of the input parameters used here are significantly uncertain and their associated error was calculated for each factory and building site. Percent uncertainty was estimated and propagated for on-site waste generation (±25%), transport distances to the modular facility (±25%), on-site ad hoc trips (±25%), and on-site temporary heating (±50%). Uncertainty associated with the choice of emissions factors, the contractor surveys, and human error in reporting is certainly present, but is not considered here.

Results and Discussion

Figure 3 shows the LCA results for GHG emissions for the three modular and five on-site companies, with averages de- picted for each construction method. The analysis reveals that impacts from modular construction are, on average, lower than those from on-site construction, but that there is significant variation within each. For example, Modular Company 1’s emis- sions were significantly higher than the other two modular cases, and also higher than one of the five on-site companies. This par- ticular facility is located in a rural area with a commute that is more than twice as long as for the other modular facilities, when normalized for production volumes. This factory also reported higher levels of electricity use than the others and was heated with fuel oil, again leading to increased levels of emissions. On

Quale et al., Construction Matters: LC A of Modular Buildings 249

R E S E A R C H A N D A N A LY S I S

T a b

le 2

A ve

ra ge

re su

lts fr

o m

m o du

la r

an d

o n-

si te

co ns

tr uc

tio n,

B EE

S 4.

02 im

pa ct

as se

ss m

en t

m et

ho d

(N IS

T 20

07 );

ro w

s m

ay no

t su

m du

e to

ro un

di ng

G H

G em

is si

on s

A ci

di fic

at io

n C

ar ci

no ge

ns N

on -c

an ce

r C

ri te

ri a

po llu

ta nt

s E

ut ro

ph ic

at io

n E

co to

xi ci

ty Sm

og W

at er

O zo

ne de

pl et

io n

(k g

C O

2 -e

q) (k

m ol

H +

) (g

be nz

en e-

eq )

(k g

to lu

en e-

eq )

(1 0−

6 D

A L

Y s)

(k g

N -e

q) (k

g 2,

4- D

-e q)

(k g

N O

x- eq

) (1

00 0

L )

(m g

C F

C -1

1- eq

)

M od

u la

r av

er ag

e 1

3 ,6

0 0

5 ,0

1 0

2 2

4 7

,8 2

0 1

,2 9

0 5

5 3

3 8

3 1

8 ,4

0 0

2 2

9 M

at er

ia lp

ro du

ct io

n 61

3 23

0 1

3, 64

0 19

1 4

3 4

1, 13

0 21

M at

er ia

lt ra

n sp

or t

to fa

ct or

y 10

3 <

1 14

< 1

1 <

1 M

od ul

e tr

an sp

or t

to si

te 1,

80 0

59 6

1 2,

59 0

86 1

1 15

24 6

1 Fa

ct or

y en

er gy

us e

6, 07

0 1,

87 0

13 28

,2 00

56 6

40 15

17 13

,9 00

25 O

n -s

it e

en er

gy us

e 3,

16 0

1, 81

0 2

5, 14

0 27

1 2

5 38

59 0

3 W

or ke

r tr

an sp

or t

to fa

ct or

y 83

1 19

3 2

2, 99

0 61

3 4

3 1,

01 0

71 W

or ke

r tr

an sp

or t

to si

te 1,

11 0

30 5

3 5,

29 0

10 7

5 5

6 1,

56 0

10 8

W as

te m

an ag

em en

t 2

1 <

1 3

7 <

1 <

1 1

< 1

O n

-s it

e av

er ag

e 1

9 ,5

0 0

7 ,7

1 0

3 0

5 8

,0 0

0 2

,1 6

0 5

1 4

7 9

5 1

5 ,5

0 0

7 4

1 M

at er

ia lp

ro du

ct io

n 78

0 27

4 2

4, 42

0 23

4 5

3 4

1, 35

0 29

M at

er ia

lt ra

n sp

or t

to fa

ct or

y 11

4 <

1 16

1 <

1 2

< 1

M od

ul e

tr an

sp or

t to

si te

0 0

Fa ct

or y

en er

gy us

e 0

0 0

0 O

n -s

it e

en er

gy us

e 11

,5 00

5, 47

0 6

19 ,4

00 1,

24 0

12 13

54 4,

01 0

17 W

or ke

r tr

an sp

or t

to fa

ct or

y 0

0 0

0 W

or ke

r tr

an sp

or t

to si

te 7,

16 0

1, 96

0 22

34 ,2

20 68

9 34

31 37

10 ,1

00 69

5 W

as te

m an

ag em

en t

3 2

< 1

5 11

< 1

1 <

N ot

es :

G H

G =

gr ee

n h

ou se

ga se

s; C

O 2 -e

q =

ca rb

on di

ox id

e eq

ui va

le n

t un

it s;

km ol

H +

= 1,

00 0

m ol

e; H

+ =

pr ot

on s;

on e

m ol

e of

pr ot

on s

(m ol

H +

); D

A L

Y =

di sa

bi li

ty ad

ju st

ed li

fe -y

ea r;

N =

n it

ro ge

n ;2

,4 -

D =

2, 4-

di ch

lo ro

ph en

ox ya

ce ti

c ac

id ;N

O x =

n it

ro ge

n ox

id es

;C FC

-1 1

= tr

ic h

lo ro

flu or

om et

h an

e. O

n e

ki lo

gr am

(k g,

S I)

= 10

3 gr

am s

(g ) =

10 6

m il

li gr

am s

(m g)

≈ 2.

20 4

po un

ds (l

b) ;o

n e

li te

r (L

) =

0. 00

1 cu

bi c

m et

er s

(m 3 ,S

I) ≈

0. 26

4 ga

ll on

s (g

al ).

250 Journal of Industrial Ecology

R E S E A R C H A N D A N A LY S I S

average, however, GHG emissions from conventional construc- tion were about 40% higher than for modular construction. Given that the production and transport of several building materials were omitted from the analysis, as they were iden- tical between construction methods, it is more appropriate to consider the absolute rather than relative difference in GHG emissions, which were found here to be nearly six metric tons of carbon dioxide equivalents (CO2-eq) higher for on-site con- struction, per 2,000 sq ft home.6 It is important to note again that this difference is likely to be modest compared to the emissions associated with building occupancy, but also that the relative importance of construction impacts will increase as the use of buildings becomes more efficient.

Differences in the materials needed for the two construc- tion methods did not significantly affect results, largely due to the fact that the additional dimensional lumber needed to give structural support to the modular panels during transport was nearly equal to the extra lumber needed to make up for wood scraps generated during on-site construction. Material transport and waste management were also small contributors to overall GHG emissions. Energy use on-site and worker transport to the site were the most important categories for GHG emissions from conventional construction, which is intuitive as both represent direct combustion of fossil fuels. Therefore, reducing unneces- sary worker trips, idling of equipment, and temporary heating through effective management practices remain the most im- portant goals of low-carbon construction of homes.

Uncertainty associated with selected input parameters is shown by the error bars of figure 3. There is moderate overlap between the sets of modular and conventionally built homes, and a much larger uncertainty range for the latter set, which reflects the sensitivity of the overall results to assumptions of temporary heating and redundant worker trips.

Table 2 shows impacts from each construction stage for the other impact assessment categories included in the BEES 4.02 methodology. The distribution of impacts among construction processes for most other impact categories is similar to that of GHG emissions: on-site construction has moderately higher impacts (20% to 70%) than modular construction, with factory and on-site energy use being the primary drivers of impacts. The exceptions to this trend are eutrophication and water intake, which are higher for modular construction due to the greater use of electricity, and ozone depletion, which is several times higher for conventional construction, due to the production and refining of crude oil needed for vehicles transporting workers and materials to the job site. This surprising result is not at all intuitive and underscores the importance of using system-level analysis tools such as LCA, as the environmental differences between two products may be due to processes far up the supply chain.

As seen in the summary rows of table 2, modular construc- tion has fewer impacts, on average, than on-site construction for all environmental impact categories, although again the uncer- tainty associated with these values is significant. Examining the three individual modular and five on-site homes individually, the Mod1 home has high impacts relative to the other modular

homes, stemming in part from its rural location, as explained above. The Conv2 home has low impacts relative to the set of conventional homes. In this particular case, the contractor who supplied the information in the study worked with a local crew and so reported relatively short distances for worker transport to the construction site, which is a major driver of impacts for most categories. This contractor also reported lower consump- tion of all fuels and electricity on-site than reported by other contractors. The variability of the responses serves to highlight the site- and company-specific nature of this study, as well as the potential for error in drawing general conclusions from single building case studies.

The environmental benefits of a compressed construction schedule for modular construction are implicit in the model inputs, as they were scaled to annual production volumes, but still bear discussion. This analysis assumes energy inputs av- eraged over the year; however, location and time of year will significantly affect the results for each of the two construc- tion methods, particularly as they will determine the number of heating-degree days needed both for the modular factory and for temporary on-site heating.

While this study has focused on residential construction, there are modular buildings in the commercial building sec- tor as well. Commercial buildings tend to have more energy- intensive structural materials (steel and concrete, as opposed to wood), therefore the additional structural material required for modular construction may have a more significant impact than for the single-family homes considered here. In addition, there is greater variety in the percentages of construction com- pleted off-site for commercial or institutional modular projects compared to single-family residential projects.

Regardless of the environmental preferences shown here, there is potential to improve environmental impacts in both methods of construction. The main opportunities for im- provement come in the energy for construction and worker transportation categories. Based on field visits and anecdotal evidence from the industry, many modular factories are the equivalent of large warehouses with little, if any, insulation. As these facilities must be heated or cooled throughout the year for worker comfort, a significant amount of this energy is wasted. For on-site homebuilders, implementation of best practices for on-site energy use (such as no idling), better mate- rial and equipment procurement policies, and implementation of carpooling could likely give some improvement. In addi- tion, it appears likely that homebuilders working on multiple homes on adjacent lots are likely to find efficiencies in material and employee transport compared to those working on discrete lots.

Acknowledgements

This research was made possible through the generosity of the many individuals in the companies that participated in the study. Partial funding for this study was provided by the Modular Building Institute, a commercial modular industry trade asso- ciation. Additional direct and indirect funding was provided

Quale et al., Construction Matters: LC A of Modular Buildings 251

R E S E A R C H A N D A N A LY S I S

by the University of Virginia (UVa) School of Architecture. UVa graduate students Rachel Lau and Emily McDermott con- tributed to the research and editing.

Notes

1. One square foot (ft2) ≈ 0.093 square meter (m2, SI). 2. One pound per square foot (lb/ft2) ≈ 4.88 kilograms per square

meter (kg/m2). 3. One mile (mi) ≈ 1.609 kilometers (km, SI). 4. One metric ton (t) = 103 kilograms (kg, SI) ≈ 1.102 short tons. 5. One megajoule per kilogram (MJ/kg) ≈ 429.9 British thermal units

per pound (Btu/lb). 6. CO2-eq: Carbon dioxide equivalent (CO2-eq) is a measure for de-

scribing the climate-forcing strength of a quantity of greenhouse gases using the functionally equivalent amount of carbon dioxide as the reference.

References

Adalberth, K. 1997. Energy use during the life cycle of single-unit dwellings: Examples. Building and Environment 32(4): 321–329.

Bilec, M., R. Ries, H. S. Matthews, and A. L. Sharrard. 2006. Example of a hybrid life-cycle assessment of construction processes. Journal of Infrastructure Systems 12(4): 207–215.

Britto, J., N. Dejonghe, M. Dubuisson, and K. Schmandt. 2007. Busi- ness plan for green modular housing. Master’s thesis, Donald Bren School of Environmental Science & Management, University of California, Santa Barbara, CA, USA.

Cole, R. J. 1998. Energy and greenhouse gas emissions associated with the construction of alternative structural systems. Building and Environment 34(3): 335–348.

Cole, R. J. and P. C. Kernan. 1996. Life-cycle energy use in office buildings. Building and Environment 31(4): 307–317.

Diez, R., V. M. Padrón, M. Abderrahim, and C. Balaguer. 2007. AUT- MOD3: The integration of design and planning tools for au- tomatic modular construction. International Journal of Advanced Robotic Systems 4(4): 457–468.

DOE (Department of Energy). 2008. Energy efficiency trends in residential and commercial buildings. Washington, DC: U.S. Department of Energy.

DSM Environmental Services. 2008. 2007 Massachusetts construction and demolition debris industry study. Boston, MA, USA: Mas- sachusetts Department of Environmental Protection.

EPA (Environmental Protection Agency). 2009. Estimating 2003 building-related construction and demolition materials and amounts. Washington, DC: U.S. EPA

EPA. 2011. Energy Star. Washington, DC: U.S. EPA. (www.energystar.gov).

Gangolells, M., M. Casals, S. Gasso, N. Forcada, X. Roca, and A. Fuertes. 2009. A methodology for predicting the severity of envi- ronmental impacts related to the construction process of residen- tial buildings. Building and Environment 44(3): 558–571.

Guggemos, A.A. and A. Horvath. 2005. Comparison of environmental effects of steel- and concrete-framed buildings. Journal of Infras- tructure Systems 11(2): 93.

Gustavsson, L. and A. Joelsson. 2010. Life cycle primary energy analysis of residential buildings. Energy and Buildings 42(2): 210–220.

Hallahan Associates. 2011. Personal communication with Fred Hallahan, Principal, Hallahan Associates. May 17, 2011; http://www.hallahanassociates.com/

Itard, L. and G. Klunder. 2007. Comparing environmental impacts of renovated housing stock with new construction. Building Research & Information 35(3): 252–267.

Junnila, S. and A. Horvath. 2003. Life-ctcle environmental effects of an office building. Journal of Infrastructure Systems 9(4): 157–166.

Keoleian, G. A., S. Blanchard, and P. Reppe. 2000. Life-cycle energy, costs, and strategies for improving a single-family house. Journal of Industrial Ecology 4(2): 135–156.

Kim, D. 2008. Preliminary life cycle analysis of modular and con- ventional housing in Benton Harbor, Michigan. Master’s the- sis, School of Natural Resources and Environment, University of Michigan, Ann Arbor, MI, USA.

Kohler, N. and S. Moffatt. 2003. Life-cycle analysis of the built envi- ronment. UNEP Industry and Environment 26(2): 17–21.

Malin, N. 2005. Life cycle assessment for whole buildings: Seeking the holy grail. Building Design & Construction (November): 6–11.

NAHB (National Association of Home Builders). 1997. Residen- tial construction waste management, a buyer’s field guide: How to save money and landfill space. Upper Marlboro, MD, USA: NAHB.

Nishioka, Y., Y. Yanagisawa, and J. D. Spengler. 2000. Saving energy versus saving materials: Life-cycle inventory analysis of housing in a cold-climate region of Japan. Journal of Industrial Ecology 4(1): 119–135.

NIST (National Institute of Standards and Technology). 2007. Build- ing for environmental and economic sustainability (BEES). Gaithersburg, MD, USA: NIST, Building and Fire Research Lab- oratory.

NREL (National Renewable Energy Laboratory). 2011. U.S. Life Cycle Inventory. Golden, CO: U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, NREL.

Ochoa, L., C. Hendrickson, and H. S. Matthews. 2002. Economic input-output life-cycle assessment of U.S. residential buildings. Journal of Infrastructure Systems 8(4): 132–138.

Oregon Department of Environmental Quality. 2010. A life cycle ap- proach to prioritizing methods of preventing waste from the residential construction sector in the state of Oregon. Phase 2 full report, 10- LQ-022. Portland, OR.

Ortiz, S., F. Castells, and G. Sonnemann. 2009. Sustainability in the construction industry: A review of recent developments based on LCA. Construction Building Materials 23(1): 28–39.

Pérez-Lombard, L., J. Ortiz, and C. Pout. 2008. A review on buildings energy consumption information. Energy and Buildings 40(3): 394– 398.

Priemus, H. 2005. How to make housing sustainable? The Dutch ex- perience. Environment and Planning B: Planning and Design 32(1): 5–19.

Scheuer, C., G. A. Keoleian, and P. Reppe. 2003. Life cycle energy and environmental performance of a new university building: Modeling challenges and design implications. Energy and Buildings 35(10–11): 1049–1064.

Sharrard, A. L. 2007. Greening construction processes using an input- output-based hybrid life cycle assessment model. Doctoral thesis, Department of Civil and Environmental Engineering, Carnegie Mellon University, Pittsburgh, PA.

Smith, R. 2010. Prefab architecture: A guide to modular design and con- struction. New York: Wiley.

252 Journal of Industrial Ecology

R E S E A R C H A N D A N A LY S I S

Treloar, G. J., P. E. D. Love, O. O. Faniran, and U. Iyer-Raniga. 2000. A hybrid life cycle assessment method for construction. Construction Management and Economics 18(1): 5–9.

About the Authors

John Quale is an associate professor at the University of Virginia School of Architecture, Charlottesville, VA, USA, where he directs the Graduate Architecture program. Matthew Eckelman is an assistant professor in the Department of Civil & Environmental Engineering at Northeastern University,

Boston, MA, USA. Kyle Williams was a masters student at the School of Forestry and Environmental Studies at Yale University at the time of writing, and is currently enrolled in the Stanford Design Program, Stanford, CA, USA. Greg Sloditskie is a principal at MBS Consulting, Inc., a consul- tancy focused on modular buildings, located in New York, NY, USA and West Milton, PA, USA. Julie Zimmerman is an associate professor in the Department of Chemical & En- vironmental Engineering and the School of Forestry & En- vironmental Studies at Yale University, New Haven, CT, USA.

Quale et al., Construction Matters: LC A of Modular Buildings 253