See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/239387890

Steel versus Steel-Reinforced Concrete Bridges: Environmental Assessment

Article  in  Journal of Infrastructure Systems · September 1998

DOI: 10.1061/(ASCE)1076-0342(1998)4:3(111)

CITATIONS READS

110 2,955

2 authors, including:

Chris Hendrickson
Carnegie Mellon University
273 PUBLICATIONS   9,789 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Transportation Data Analytics View project

https://www.nap.edu/catalog/14670/underground-engineering-for-sustainable-urban-development View project

All content following this page was uploaded by Chris Hendrickson on 17 April 2015.

The user has requested enhancement of the downloaded file.

 STEEL VERSUS STEEL-REINFORCED CONCRETE BRIDGES:
ENVIRONMENTAL ASSESSMENT

By Arpad Horvath l and Chris Hendrickson2

ABSTRACT: Bridge material selection has traditionally been b~ed on e.ngineering and economi~ criteria. With
the increasing interest of the public, industry, and government m sustainable:: development, en~lronmental. as-
sessment in construction is becoming more important. However, we need metrtcs and data on deSign alternatives
for effective decision making. In this paper, we present results of a life cycle inventory analysis of steel and
steel-reinforced concrete bridge girders, based on publicly available data. We find that for the initial constru~tion
of equivalent designs for a particular location, a steel-reinforced concrete bridge g~neral~y .has lower enViron-
mental effects than a steel bridge. The expected design life of the two types of bndges is mfluenc~d. by wear
and tear, but also by obsolescence. The uncertainty in bridge design life ~d r 1ated ~ta uncertamties m~e
comparisons based on annualized environmental effects difficult. The ~teel bndge.s7b.eneficial reuse and recych?g
rates may result in lower annualized environmental effects. In particular apphcatlOns, however, one matenal
might be preferred over the other due to engineering, aesthetic, or economic criteria, regardless of overall
environmental effects.

INTRODUCTION ficult to perform due to a lack of appropriate scientific data

Interpretation of the results has been controversial, depending
Bridge superstructure designs typically involve steel or on the objectives of the reader, and on the weight assigned to
steel-reinforced concrete, or a combination of the two mate- each of the environmental effects quantified (Portney 1994).
rials. The two designs are competitive, with numerous exam- Many LCA studies produced to date have not been published
ples of both types in use. The tra~itio~al crite~ia for sel~c~i?g due to the proprietary data use or the controversial results.
a particular design have been engmeenng reqUirements, mitial Despite its limitations, there is an articulated need for LCA
and life cycle costs, experience with and availability of a par- studies, as well as better metrics and data, in decision making
ticular material or technology, aesthetics, and the ability to for sustainable development.
erect the structure under local environmental conditions (cli- In this paper an economic input-output-based life cycle as-
mate, topography, etc.). With a steel and steel-reinforced con- sessment (EID-LCA) is used. This method is predicated on the
crete comparison, we would like to draw attention to the en- fact that in a modern economy every sector contributes, di-
vironmental implications of particular bridge material choices rectly or indirectly, to every other sector (Lave et al. 1995;
and designs. The environmental assessment of the two mate- Hendrickson et al. 1998). The EIO-LCA model employs eco-
rials is best performed using life cycle assessment (LCA). nomic input-output analysis, invented by Wassily Leontief in
the 1930s (for which he received a Nobel prize) (Leontief
LIFE CYCLE ASSESSMENT (LCA) 1986), and commodity-by-commodity input-output matrices,
LCA is a method that systematically assesses and analyzes compiled by the Department of Commerce ("Input-output"
the environmental effects of the five various stages of a prod- 1994) in the United States, and available in similar form for
uct's or process' entire life cycle: (1) the materials extraction most developed and industrializing countries. Economic input-
phase; (2) the materials processing stage; (3) the manufactur- output analysis explicitly accounts for all of the direct and
ing stage; (4) the use phase; and (5) the ultimate disposal stage indirect inputs to producing a product or service by using the
(end-oC-life). LCA has three parts: (1) the inventory analysis input-output matrices of a national economy. For example, ce-
ment production and electricity are direct inputs of concrete
quantifies the environmental effects; (2) the impact analysis
production, but many agricultural and service sectors are in-
estimates the effects of these burdens on humans and nature;
and (3) the improvement analysis identifies areas and means direct, or second, third, etc., tier suppliers (by being suppliers,
of possible improvement. There are many efforts in the world or suppliers of suppliers of direct suppliers). In addition, elec-
tricity is required to produce electricity, and cement is needed
to produce LCAs for a variety of products and processes (Vi-
for cement manufacturing. Suppose that the production of $1
gon et al. 1993; Graedel and Allenby 1995). Data are gathered
of concrete requires $0.20 of inputs from the cement industry,
either directly from participants in the life cycle chain of a
which in tum requires $0.10 of coal mining, etc. The conven-
product or process, or used from libraries of past LCAs. For
tional approach would have to limit the number of process
these reasons, LCA studies, in general, have been expensive
flows and suppliers to complete the assessment. The input-
and time-consuming. If a product is redesigned, the analysis
output model accounts for all the direct and indirect economic
has to be redone. Most LCA studies performed to date have
effects.
been, in fact, inventory analyses, with some attempt to include Environmental input-output analysis complements the eco-
improvement analysis. Impact analysis has especially been dif-
nomic input-output analysis by linking economic data with
IRes. Engr., Dept. of Civ. and Envir. Engrg., Carnegie Mellon Univ.,
resource use (such as energy, ore, and fertilizer consumption)
Pittsburgh, PA 15213; corresponding author. E-mail: ah3p@andrew. and environmental output [such as toxic chemical discharges,
cmu.edu Subtitle C hazardous waste generation and management as de-
2Duquesne Light Prof. of Engrg. and Head, Dept. of Civ. and Envir. fined in the Resource Conservation and Recovery Act (U.S.
Engrg., Carnegie Mellon Univ., Pittsburgh, PA. E-mail: [email protected] EPA 1993), ozone depletion potential, and conventional air
Note. Editor-in-Chief: Jeff R. Wright. Discussion open until February pollutant emissions] data. In the above example, total direct
I, 1999. To extend the closing date one month, a written request must
be filed with the ASCE Manager of Journals. The manuscript for this
and indirect (upstream) emissions of species i per $1 of con-
paper was submitted for review and possible publication on August 25, crete product can be calculated by taking the sum: (concrete
1997. This paper is part of the JourllQl of Infrastructure Systems, Vol. production's direct emissions of i per $1 of concrete) +
4, No.3, September, 1998. @ASCE, ISSN 1076-0342198/0003-0II 1- 0.20*(cement industry's direct emissions of i per $1 of ce-
01 17/$8.00 + $.50 per page. Paper No. 16456. ment) + 0.20*0.1O*(coal mining's emissions of i per $1 of
JOURNAL OF INFRASTRUCTURE SYSTEMS 1 SEPTEMBER 1998/111

Downloaded 14 Jun 2012 to 128.237.190.83. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
coal) + other tenns representing the inputs to coal mining, and the manufacturing stages, as if $781,000 worth of steel-
etc. In reality, there are many more inputs in each industry reinforced concrete girders, and $1,756,000 worth of steel
than described here. We have used a 519*519 commodity-by- girders were purchased (final demand). The two girder mate-
commodity matrix of the U.S. economy that identifies and rials were represented by sectors in the economic input-output
quantifies what those inputs are for U.S. industries and com- matrix. For the concrete girder, we used the sector concrete
modities. products (except block and brick) (Standard Industrial Clas-
In this article, we present a life cycle inventory analysis of sification, SIC, code 3272). For the steel girder, the fabricated
steel and steel-reinforced concrete bridge girders. We used the structural metal sector (SIC 3441) was used. Producing the
EIO-LCA method to assess the environmental effects of the concrete girders generated an intennediate demand (in input
materials extraction, the materials processing, and the manu- products and services) of $686,000 in the economy (for a total
facturing stages, and data from literature for the other life cy- demand of $781,000 + $686,000 = $1,467,000), and produc-
cle phases. This article is based on an unpublished doctoral ing the steel girders had an intennediate demand of $2,142,000
dissertation (Horvath 1997). (for a total demand of $3,898,000).
Several resource input requirements and environmental bur-
DESIGN AND MATERIAL COMPOSITION OF BRIDGE dens have been quantified. Resource inputs include consump-
GIRDERS tion of electricity, fuels, ores, and fertilizers ("Input-output"
1994). Fertilizers are not direct inputs to either steel or con-
Alternative designs for the bridge on U.S. 231 over the crete manufacturing, but they are part of the long chain of
White River in Indiana (Spaans 1997) were selected. Bids indirect suppliers, that is, upstream suppliers of direct suppliers
were solicited for a 428.2 m (l,405 ft) long and 14.7 m (48.3 such as forestry products. Environmental outputs include toxic
ft) wide bridge with either steel or reinforced concrete girders. chemical discharges to air, water, land, underground injection
We used the data for the low bids using steel plate girders and wells, and transfers to off-site treatment plants (U.S. EPA
posttensioned concrete girders. Table 1 presents the cost fig- 1995), ozone depletion potential of chemical releases (U.S.
ures for the two designs. A concrete deck was designed for EPA 1995); Ozone 1997-website: http://www.epa.gov/
both the steel and the concrete girder bridge, and the cost of ozone/title6/procu/procu.html), hazardous waste generation
the deck and the cross girders (with the rebar in deck and cross and management (U.S. EPA 1993), and conventional air pol-
girders) in both cases were comparable. It is evident from Ta- lutant emissions ("Input-output" 1994; "Air CHIEF" 1995).
ble 1 that the cost of the substructure for the two designs was Similar to the resource input requirements, the environmental
close, with the concrete bridge 13% less expensive. There was effects of not only the direct suppliers (such as the cement
a difference in the cost of the expansion joints ($53,000 for industry for concrete), but the indirect suppliers as well (such
the steel design and $35,000 for the concrete design) and in as the agricultural sector for concrete) are included in this as-
the cost of bearings ($94,000 for the steel design versus sessment.
$20,000 for the concrete design). Only the girders were com-
pared since the difference between the two designs is mostly UNCERTAINTY IN DATA
due to the difference in the cost of girder materials.
All of the data used in this study are uncertain. For example,
The steel plate girder bridge design had eight spans of var-
concrete bridge girders were estimated by a sector (concrete
ious lengths, from 27.4 to 54.9 m (90 to 180 ft), and two
products) that include not only girders but other products, with
additional spans totaling 61.7 m (202.5 ft) made with
perhaps different environmental implications. Furthennore, the
AASHTO Type IV girders. The posttensioned bulb-tee girder
toxic chemical releases' data are obtained from the EPA's Tox-
design had nine equal spans of 40.8 m (133.75 ft) and two ics Release Inventory (TRI), collected from manufacturing
additional spans made with the same type and length of girders
plants in the United States. Facilities have to report to the TRI;
as the steel alternative. Since the two additional spans were
they do not have to measure their emissions but do have to
identical, we neglected them in our analysis. Overall, the con- estimate. This inevitably leads to inaccuracies (U.S. GAO
crete alternative had a $1,000,000 lower initial cost (a 37%
1991). Similar uncertainties exist regarding RCRA Subtitle C
savings). hazardous waste data. Conventional pollutant emissions are
based on the fuel consumption of facilities, and the AP-42
PERFORMING EIO-LCA emissions factors ("Air CHIEF" 1995) from the EPA. The
The BIO-LCA method was used to assess the environmental AP-42s are estimates of air emissions from processes. Re-
effects of the materials extraction, the materials processing, source input calculations are based on data gathered by the
U.S. Department of Commerce. Companies' own accounts of
TABLE 1. Low Bids for U.S. 231 over the White River In Indiana these costs are not perfect, and many facilities fail to report
(Spaans 1997) them completely. Therefore, the results obtained in this LCA
must be analyzed in light of the uncertainty in the data.
Design Cost
(1 ) (2) RESOURCE INPUT REQUIREMENTS
Steel alternative Table 2 contains a summary of the resource input require-
Steel girders + bearings 1,756,000 +
93,870
Expansion joints 52,795 ments associated with materials extraction, materials process-
Concrete deck and cross girders + rebar 565,628 + 315,748 ing, and manufacturing steel and steel-reinforced concrete
Substructure 855,039 girders for the highway bridge in our example. With the ex-
Bridge, total 3,639,080 ception of some fertilizer consumption, all other resource in-
Cost per square foot of steel bridge 53.59 puts appear to be higher for the steel alternative. If we convert
Concrete alternative fuel usage by type into a common unit, it is roughly 12,000
Concrete girders + bearings 781,000 +
20,000
Post-tensioning + erection 135,000 +
75,000 GJ of energy for the concrete and 33,000 GJ of energy for the
Expansion joints 34,720 steel design.
Concrete deck and cross girders + rebar 575,009 + 281,831
Substructure 747,191 ENVIRONMENTAL OUTPUTS
Bridge, total 2,649,751
Table 3 contains a summary of the outputs associated with
Cost per square foot of concrete bridge 39.02
the materials extraction, materials processing, and manufac-
112/ JOURNAL OF INFRASTRUCTURE SYSTEMS / SEPTEMBER 1998

Downloaded 14 Jun 2012 to 128.237.190.83. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
TABLE 2. Environmental Effects of Steel-Reinforced Con- TABLE 3. Environmental Effects of Steel-Reinforced Con-
crete (Concrete + Reinforcing) and Steel Production for Exam- crete (Concrete + Reinforcing) and Steel Production for Exam-
ple Equivalent Bridge Girder Design-Summary of Resource ple Equivalent Bridge Girder Design-Summary of Environ-
Inputs mental Outputs
Ratio Ratio
(concrete/ (concrete/
Resource Inputs Concrete Steel steel) Environmental outputs Concrete Steel steel)
(1 ) (2) (3) (4) (1) (2) (3) (4)
Electricity (kW' h million) 0.4 1.4 0.3 TRI air releases (t) 0.1 0.7 0.1
Coal and coke TRI water releases (t) 0.02 0.07 0.3
Anthracite coal (t) 0.5 1.4 0.4 TRI land releases (t) 0.06 0.6 0.1
Bituminous coal (t) 300 800 0.4 TRI underground releases (t) 0.04 0.2 0.2
Coke (t) 8 100 0.1 TRI total releases to environment (t) 0.2 1 0.2
[Total] (t) 309 901 0.3 TRI total releases and transfers (t) 0.8 7 0.1
Fuels CMU-ET for air releases (t H,S04
Natural gas (t) 20 60 0.3 equivalent) 0.03 0.3 0.1
Liquefied natural gas (t) 3 5 0.6 CMU-ET for water releases (t H,S04
Motor gasoline (t) 10 20 0.5 equivalent) om 0.05 0.2
Aviation fuel (t) 0.1 0.2 0.5 CMU-ET for land releases (t H,S04
Jet fuel (t) 2 7 0.3 equivalent) 0.3 3 0.1
Kerosene (t) 0.004 0.005 0.8 CMU-ET for underground releases
Light fuel oil (t) 30 50 0.6 (t H,S04 equivalent) 0.02 0.1 0.2
Heavy fuel oil (t) 7 20 0.4 CMU-ET for total releases (t H,S04
Liquefied petroleum gas (t) 6 10 0.6 equivalent) 0.4 3 0.1
[Total] (t) 78 172 0.5 CMU-ET for releases and transfers
Ores (t H,S04 equivalent) 2 20 0.1
Iron ore (t) 70 900 0.08 Ozone depletion potential (t CFC-ll
Ferrous ore (dollars) 100 700 0.1 equivalent) 0.002 0.008 0.3
Copper ore (t) 20 250 0.08 RCRA hazardous waste generated (t) 20 70 0.3
Bauxite (t) 2 8 0.3 RCRA hazardous waste managed (t) 20 40 0.5
Gold ore (t) 30 250 0.1 RCRA hazardous waste shipped (t) 3 30 0.1
Silver ore (t) 5 50 0.1 SO, (t) 10 30 0.3
Lead-zinc ore (dollars) 4 300 0.01 NO. (t) 4 10 0.4
Uranium-vanadium ore (dollars) 20 300 0.07 Methane (t) 0.03 0.05 0.6
Fertilizers Volatile organic compounds (t) 0.4 0.8 0.5
Ammonia (t) I 2 0.5 Note: Numbers may not sum due to rounding.
Ammonium nitrate (t) 1 0.5 2
Ammonium sulfate (t) 0.02 0.04 0.5
Urea (t) 0.1 0.1 1 USE PHASE ENVIRONMENTAL EFFECTS
Organic fertilizers (t) 0.01 om 1
Superphosphate (t) 0.3 0.4 0.8 Maintenance operations during the lifetime of a bridge have
Phosphatic fertilizers (t) 0.001 0.001 1 environmental consequences. Unfortunately, maintenance
[Total] (t) 2.4 3.0 0.8 costs are difficult to obtain. In this assessment only the paint-
ing of the steel structure, as perhaps the most important main-
tenance need for a steel bridge, has been considered (Rainer
turing stages of steel and concrete bridge girders. Three major 1990). We estimated that -6,040 m 2 (65,000 ff) of steel girder
groups of environmental impacts are quantified in this assess- surface needs to be coated for the steel design alternative in
ment: (l) TRI chemical emissions; (2) hazardous waste gen- the current example, and that only one coating is required.
eration; and (3) conventional air pollutant emissions. For the (Note that the initial painting of the girders and the painting
TRI discharges, we assessed both the amounts of TRI emis- of railings was not assessed above.) A catalog ("Means"
sions as reported by facilities to the EPA, without regard to 1987) lists a unit material price of $0.10 per ft 2 or $1.08 per
the relative toxicity of the emissions, and the amounts of TRI m2 for bridge repainting. It was assumed that bridge paint is
emissions weighted by relative toxicity. The latter method was produced by the paints and allied products sector (SIC 285),
developed to account for the fact that a kilogram of a TRI and that we need to purchase $6,500 worth of paint for a single
discharge may be more or less environmentally damaging than job. A maintenance operation such as bridge repainting is re-
another kilogram of a different chemical release (Horvath et peated several times during the lifetime of a bridge. It is es-
al. 1995). For example, a kilogram of lead emitted to the air timated (Rainer 1990) that a typical preventive maintenance
is more toxic than an equal amount of methanol. The weight- plan for a large steel bridge requires a repaint every eight
ing of TRI chemicals is perfonned using the threshold limit years. This is applicable to our steel girder bridge example. If
values (TLVs), which represent a maximum concentration we assume that the steel bridge will last up to 80 years, then
level in air to which workers may be exposed over a typical it will be repainted 10 times, including the first painting.
work week without adverse health effects. Although not all It is difficult to assess the relative magnitude of inputs and
TLVs are good measures of toxicity and do not account for outputs for the paint without comparing them to another ex-
the different fate and transport of substances in the environ- ample. Here we compare the paint figures to the inputs and
ment, they are available for the greatest number of TRI chem- outputs of steel bridge girders. Table 4 compares the resource
icals and are a reasonable first approximation of human tox- inputs. Table 5 compares the environmental outputs of the
icity. We call the weighting method Carnegie Mellon manufacture of the steel girders and the paint for ten repaint
University-Equivalent Toxicity (CMU-En, and the weighted jobs for the steel bridge. Note that, as expected, with the ex-
pounds of emissions are nonnalized to kilograms of sulfuric ception of conventional air pollutants, all resource require-
acid equivalent. ments and environmental outputs are at least several times
After examining the environmental outputs of concrete and higher for the girder production. Of course, if the bridge would
steel girder manufacturing, we find that emissions appear to be repainted more often and in more layers, the differences in
be higher for the steel girders. the numbers would be smaller. However, Table 5 shows a sur-
JOURNAL OF INFRASTRUCTURE SYSTEMS / SEPTEMBER 1998/113

Downloaded 14 Jun 2012 to 128.237.190.83. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
TABLE 4. Environmental Effects of Paint (for Ten Repaint TABLE 5. Environmental Effects of Paint (for Ten Repalnl
Jobs) and Steel Manufacturing for 'TYpical Highway Brldge- Jobs) and Steel Manufacturing for 'TYpical Highway Brldge-
Summary of Resource Inputs Summary of Environmental Outputs
Ratio Ratio
Resource Inputs Paint Steel (paint/steel) (paint/
(1 ) (2) (3) (4) Environmental outputs Paint Steel steel)
Electricity (kW' h million) 0.04 1.4 0.03 (1 ) (2) (3) (4)
Coal and coke TRI air releases (t) 0.06 0.7 0.09
Anthracite coal (t) 0.02 1.4 0.01 TRI water releases (t) 0.008 0.07 0.1
Bituminous coal (t) 8 800 0.01 TRI land releases (t) 0.008 0.6 0.01
Coke (t) 0.4 100 0.004 TRI underground releases (t) 0.05 0.2 0.3
[Total] (t) 8 901 0.009 TRI total releases to environment (t) 0.2 1 0.2
Fuels TRI total releases and transfers (t) 0.4 7 0.06
Natural gas (t) 3 60 0.05 CMU-ET for air releases (t H2SO.
Liquefied natural gas (t) 0.3 5 0.06 equivalent) 0.008 0.3 0.03
Motor gasoline (t) 0.3 20 0.02 CMU-ET for water releases (t H2SO4
Aviation fuel (t) 0.004 0.2 0.02 equivalent) 0.003 0.05 0.06
Jet fuel (t) 0.2 7 0.03 CMU-ET for land releases (t H2SO4
Kerosene (t) 0.0002 0.005 0.04 equivalent) 0.06 3 0.02
Light fuel oil (t) 2 50 0.04 CMU-ET for underground releases
Heavy fuel oil (t) 0.5 20 0.03 (t H2S04 equivalent) 0.03 0.1 0.3
Liquefied petroleum gas (t) 0.3 10 0.03 CMU-ET for total releases (t H2SO4
[Total] (t) 7 172 0.04 equivalent) 0.08 3 0.03
Ores CMU-ET for releases and transfers
Iron are (t) 2 900 0.002 (t H2S04 equivalent) 0.2 20 0.01
Ferrous are (dollars) 20 700 0.03 Ozone depletion potential (CFC-ll
Copper ore (t) 3 250 0.01 equivalent) 0.002 0.008 0.3
Bauxite (t) 3 8 0.4 RCRA hazardous waste generated (t) 8 70 0.1
Gold ore (t) 2 250 0.008 RCRA hazardous waste managed (t) 3 40 0.08
Silver ore (t) 0.4 50 0.008 RCRA hazardous waste shipped (t) 0.8 30 0.03
Lead-zinc are (dol1ars) 20 300 0.07 S02 (t) 200 30 67
Uranium-vanadium ore (dol1ars) 2 300 0.007 NO, (t) 80 10 8
Fertilizers Methane (t) 0.8 0.05 16
Ammonia (t) 0.4 2 0.2 Volatile organic compounds (t) 20 0.8 25
Ammonium nitrate (t) 0.02 0.5 0.04
Ammonium sulfate (t) 0.006 0.06 0.2
Urea (t) 0.02 0.1 0.2
Organic fertilizers (t) 0.002 0.01 0.2 The U.S. statistics on steel product shipments (AMM 1996)
Superphosphate (t) 0.06 0.40 0.2
Phosphate fertilizers (t) 0.0001 0.001 0.1 are broken down by construction uses, such as structural
[Total] (t) 0.5 3 0.2 shapes. However, steel bridge girders are uniquely designed
and manufactured products, welded and bolted together from
steel plates. Steel plates, on the other hand, are used not only
in construction, but in automotive applications, in oil and gas
prising result. Sulfur dioxide, oxides of nitrogen, methane, and industry, in mechanical equipment manufacturing, in ship-
volatile organic compound emissions are significantly higher building, etc. Therefore, it is difficult to determine how much
for the paint manufacturing than for the production of all gird- steel plate is used in bridge applications on a yearly basis.
ers for the example highway bridge. Therefore, the environ- Due to the lack of comprehensive data on national steel
mental effects of the use phase of products can be very im- statistics, data from a joint Federal Highway Administration
portant in life cycle assessment. (FHWA)-EPA study (Bloomquist et at. 1993) were used to
estimate the recycling rates for steel girders. This report con-
END-OF-LIFE OPTIONS FOR BRIDGE GIRDERS tains results from a survey of the recycling practices of 29
state highway agencies in the United States conducted at the
Steel Girders end of 1992. Data were collected on bridge superstructures,
Le., on beams and decks, not solely on steel structural mem-
Steel bridge girders last a long time. Some steel bridges bers. Therefore, the results could be skewed if bridge decking
constructed in the last century still survive with regular main- reuse or recycling rates had been different from those for the
tenance and repair. The decommissioning of steel bridges is girders. The results of the survey indicate that reuse of obso-
often the practice not because the girders reach the end of their lete bridge steel superstructures has been practiced by 20 out
structural life, but because of functional obsolescence. Chang- of 29 states. Rates of reuse range from 1% in Maryland and
ing traffic volumes, loads, or patterns may require a wider, Wyoming to 100% in Vermont, with 13 states between 1 and
stronger, larger, longer bridge (Lerner 1996). Often the major 20%, and six states above 40%. However, there is an overlap
traffic routes move away from the bridge. Especially in remote between reuse and recycling in this report. Recycling of steel
areas, or where historic preservation efforts have saved them, superstructures in this context meant cutting, breaking, or
many steel bridges have not been deconstructed, but left in modifying the steel superstructure for use in a different high-
place, closed for traffic. In some instances, with smaller way application, or reusing or storing it for subsequent use
bridges, the superstructure has been reused at another location after straightening, painting, or minor repair. Nine states re-
where the old bridge structure was sufficient for the local traf- ported recycling rates between 5%, in Arizona, and 100%, in
fic. This represents a beneficial reuse of steel girders. Since Utah, with seven states achieving more than a 50% recycling
they can be a feedstock for new steel production, it is pre- rate. Disposal options included sales as scrap, landfilling, giv-
sumed that, if not left in place or reused, obsolete steel bridge ing the superstructure to a contractor or to others, and dispos-
girders are recycled. Comprehensive, national data on steel ing of unusable or unsuitable items. Therefore, even though
bridge girder recycling in the United States are unavailable. the disposal rates were the highest of all three end-of-initial-
114/ JOURNAL OF INFRASTRUCTURE SYSTEMS / SEPTEMBER 1998

Downloaded 14 Jun 2012 to 128.237.190.83. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
life options for 21 states out of the 29, it did not automatically years), therefore, they are comparable. Steel bridges have been
mean 1andfilling. Eleven states explicitly noted that the obso- constructed for a longer period of time than concrete bridges.
lete steel became the property of the contractor who disman- As a result of regular maintenance, extensive repair and mod-
tled it, who in turn could have reused, recycled, or eventually ifications, some parts of steel bridges can survive for 100 years
landfilled it. One state, Connecticut noted that only unusable or more. The first prestressed concrete bridge was not finished
items were landfilled. Three states sold the obsolete steel as until 1951 (Marland and Weinberg 1988), and it was the build-
scrap. Therefore, the reuse and recycle rates in Bloomquist et ing of the Interstate highway system that brought about the
al. (1993) (17 and 21 %, respectively) are actually minimum prevalence of concrete highway bridges in the U.S. Of course,
rates. For example, Virginia reported that 100% of the dis- six factors other than time may also influence the useful life
mantled steel superstructure was disposed of, but noted that it of bridges: (1) flood; (2) fire; (3) wind; (4) foundation scour;
was sold as scrap and therefore rec' ·;;rt:d. For the same state, (5) war; and (6) collision (Marland and Weinberg 1988). Most
the recycling rate was reported as zero, when in effect it was importantly, with everincreasing traffic and changing societal
100%. Data for five Canadian provi:.ces are also provided in demands, functional obsolescence, not time, might render any
the report, exhibiting a similar trend to the U.S. states. type of bridge obsolete long before it fails structurally (Lemer
Of course, these survey results should be interpreted care- 1996).
fully. There might have been significant reporting errors, and How many resources are embedded in the steel bridges, and
it was unspecified in the survey what period the report had what percentage of the national emissions does the manufac-
covered (e.g., 1992 only, last three years, or standard practice). turing of steel bridge girders account for? Of the currently
Given that only a small number of bridges are decommis- 580,000 bridges in the United States, we do not know exactly
sioned and deconstructed every year in individual states, the how many have steel versus concrete superstructures. We may
survey might not be representative of actual time trends. assume that half of the bridges (290,000) were steel, and on
average they required about $1,800,000 worth of steel girders.
Concrete Girders Table 6 summarizes the percentages of the U.S. national totals
that resource inputs into manufacturing steel girders for
Similarly to the steel design alternative, the best source of 290,000 average bridges accounted for. Table 7 shows the per-
statistics on concrete girder recycling is the joint FHWA-EPA centages for environmental outputs. The largest percentage be-
study of 1993 (Bloomquist et al. 1993). However, the end-of- longs to iron ore. If half of the U.S. bridges were steel, the
life options for concrete beams seem to be much more limited iron ore consumption would amount to more than 5 years'
than for steel beams. Of the 27 states responding to the survey aggregate national demand. Similarly, bauxite consumption
in this study, only six reported any reuse of old concrete gird- would amount to 3 years', and coke consumption would
ers, with the rates ranging from 10 to 50%. Recycling has only amount to 2 years' U.S. demand. Building 290,000 new steel
been reported by four states at a 10, 50, 70, and 100% rate,
respectively. Consequently, a large majority of the states re- TABLE 6. Re80urcelnputs for Manufacturing Steel Girders for
ported a 100% disposal rate, with the option of the old con- 290,000 Average Bridges, as Percentage of Annual National Ta-
crete beams going into landfills or being given to contractors. tals In the United States
The contractors may have reused or recycled the old concrete Annual national
beams, thus, raising their reuse and recycling rates, or they For 290,000 total
may have eventually landfilled them, for lack of a better use. Resource inputs bridges (%)
Therefore, as with the steel alternative, the reported reuse and (1 ) (2) (3)
recycling rates may be underestimated. Upon examining the Electricity (kW, h million) 406,000 17
numbers for five Canadian provinces, it is noted that no reuse Coal and coke
and, in only one province, was recycling observed. Of course, Anthracite coal (t) 406.000 9
data and survey quality issues might make these numbers un- Bituminous coal (t) 232,000,000 39
reliable or unrepresentative, as with the steel girder option. Coke (t) 29.000.000 239
Unrepresentativeness is possible, given the large differences in Natural gas (t) 17,400.000 21
the reported practices: (1) two states observed a 50% reuse Liquefied natural gas (t) 1,450.000 7
Motor gasoline (t) 5,800.000 2
rate, and (2) one state reported a 100% recycling rate when Aviation fuel (t) 58.000 2
many other states might have landfilled their old concrete Jet fuel (t) 2,030,000 4
beams entirely. Kerosene (t) 1,450 0.06
Light fuel oil (t) 14,500.000 12
DISCUSSION Heavy fuel oil (t) 5.800,000 13
Liquefied petroleum gas (t) 2.900.000 10
Some environmental effects associated with the materials Ores
Iron ore (t) 261.000.000 548
extraction, materials processing, and manufacturing stages of Ferrous ore (dollars) 203,000,000 92
steel-reinforced concrete and steel bridge girders for equiva- Copper ore (t) 72,500.000 36
lent designs are presented previously. The concrete design ap- Bauxite (t) 2,320,000 287
pears to have lower environmental effects overall. Of course, Gold ore (t) 72.500.000 81
the results might be different for another design as every Silver ore (t) 14,500,000 119
bridge is unique in its design and material content. Lead-zinc ore (dollars) 87.000.000 33
Uranium-vanadium ore (dollars) 87,000.000 35
The bridge girders were compared based on a summary of Fertilizers
environmental effects. However, there might be a difference Ammonia (t) 580,000 6
in the expected design life of the two materials. For a more Ammonium nitrate (t) 145.000 4
realistic comparison, we need to take into account longevity Ammonium sulfate (t) 11.600 0.6
and annualize environmental effects. However, comprehensive Urea (t) 29.000 0.6
statistics on the expected life of steel and concrete bridges are Organic fertilizers (t) 2,900 0.2
Superphosphate (t) 116.000 2
hard to find. One source (Veshosky and Nickerson 1993) es- Phosphatic fertilizers (t) 290 0.2
timates the life of bridges in Belgium, Japan, Sweden, and
Switzerland at 47-76 years for steel, and 47-86 years for Note: Resource inputs and corresponding national totals are based on
1987 data.
reinforced concrete (prestressed concrete bridges last 21- 86
JOURNAL OF INFRASTRUCTURE SYSTEMS / SEPTEMBER 1998/115

Downloaded 14 Jun 2012 to 128.237.190.83. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
TABLE 7. Environmental Outputs for Manufacturing Steel TABLE 8. Summary of Environmental Effects of Steel versus
Girders for 290,000 Average Bridges as Percentage of Annual Steel-Reinforced Concrete Girders for lYplcal Highway Bridge
National Totals In the United States
Type of
Annual national girder Manufacturing Use End-of-Iife option
For 290,000 total (1 ) (2) (3) (4)
Environmental outputs bridges (%) Steel Likely higher resource Paint, other Better reuse and
(1) (2) (3) input requirements maintenance recyclability
TRI air releases (t) 203,000 32 and environmental
TRI water releases (t) 20,000 22 outputs
TRI land releases (t) 174,000 140 Concrete Better Other mainte- Mostly landfilling
TRI underground releases (t) 58,000 43 nance
TRI releases to environment (t) 290,000 30
TRI total releases and transfers (t) 2,030,000 83
CMU-ET for air releases (t H,S04 in the case of visual impacts, the lack of an acceptable metric.
equivalent) 87,000 64 For example, dust emissions, water usage, nonhazardous solid
CMU-ET for water releases (t H,S04 waste generation and disposal, generation and disposal of haz-
equivalent) 14,500 70
CMU-ET for land releases (t H,S04
ardous waste by type, environmental effects of landfills, noise
equivalent) 870,000 109 and vibration, and visual impacts have not been assessed. If
CMU-ET for underground releases these (and other) environmental effects would have been in-
(t H,S04 equivalent) 29,000 70 cluded, our assessment might have yielded different conclu-
CMU-ET for total releases (t H,S04 sions. Also, the data used in this analysis have large uncer-
equivalent) 870,000 79 tainties associated with them, and they reflect past economic
CMU-ET for releases and transfers
(t H,S04 equivalent) 5,800,000 66
and environmental performance. Therefore, a similar assess-
Ozone depletion potential (t CFC-ll ment using different designs and baseline years may yield dif-
equivalent) 2,320 14 ferent conclusions.
RCRA hazardous waste generated (t) 20,300,000 9 If, however, obsolescence is a main problem, it might not
RCRA hazardous waste managed (t) 11,600,000 5 matter if one material lasts longer than the other because a
RCRA hazardous waste shipped (t) 8,700,000 55 bridge might be decommissioned long before it fails structur-
SO, (t) 8,700,000 40
NO, (t) 2,900,000 35
ally. If indeed this might be the case, bridges should be built
Methane (t) 14,500 40 from the material that has comparably the lowest environmen-
VOC (t) 232,000 50 tal burdens. In particular applications, however, engineering,
Note: Environmental outputs and corresponding national totals are
aesthetic, or economic criteria might outweigh the environ-
based on 1993 data. mental factor.

ACKNOWLEDGMENTS
bridges in the United States would raise considerably environ-
The writers wish to thank the National Science Foundation's Struc-
mental outputs as well. Since steel bridges in the United States tures, Geomechanics and Building Systems program (Project CMS 97-
have not been built in 1 year, but over many decades, the 13917), and the Department of Energy's Office of Health and Environ-
annualized input and output totals attributed to bridges would mental Research for financial support for this work. Profs. Lester Lave
be small. The comparison to annual national totals is insightful and Francis McMichael from Carnegie Mellon University provided help-
as we do not regularly think of bridges as sinks of renewable ful comments.
and nonrenewable resources and direct and indirect causes of
pollution. APPENDIX. REFERENCES
Air CHIEF. (1995). CD-ROM, U.S. Envir. Protection Agency, Washing-
CONCLUSIONS ton, D.C.
When looking at the initial construction of equivalent bridge American Metal Market. (AMM). (1996). Metal statistics I996-ferrous
designs, steel-reinforced concrete girders appear to have lower edition. The Statistical Guide to North American Metals. 88th Ed.,
Chilton Publications, New York, N.Y.
overall environmental effects than steel girders. However, steel Bloomquist, D., Diamond, G., Oden, M., Ruth, B., and Tia, M. (1993).
girders are reusable and recyclable at the end of their useful "Engineering and environmental aspects of recycled materials for high-
life. Steel superstructures have had a documented minimum way construction-Final Report," Rep. No. FHWA-RD-93-088, U.S.
reuse rate of 17% and a minimum recycle rate of 21 %, based Dept. of Transp., Fed. Hwy. Admin., and U.S. Envir. Protection
on a limited U.S. national survey (Bloomquist et al. 1993). Agency.
For concrete girders, the options have mostly been limited to Graedel, T. E., and Allenby, B. R. (1995). Industrial ecology. Prentice-
Hall, Inc., Englewood Cliffs, N.J.
landfilling in the past. The reuse and recycle rates for steel Hendrickson, C., Horvath, A., Joshi, S., and Lave, L. (1998). "Economic
girders presumably save input resources and reduce environ- input-output models for environmental life cycle assessment." Envir.
mental pollution compared to using only virgin materials. Re- Sci. and Technol., April, 184A-191A.
cycling rates for steel girders are taken into account in EIO- Horvath, A. (1997). "Estimation of the environmental implications of
LeA through the current mix of raw materials for steel mills. construction materials and designs using life cycle assessment tech-
In 1987, roughly 20-30% of steel was recycled. However, niques," PhD thesis, Dept. of Civ. and Envir. Engrg., Carnegie Mellon
Univ., Pittsburgh, Pa.
Malin (1997) reports that steel plates used to fabricate struc- Horvath, A., Hendrickson, C., Lave, L. B., McMichael, F. C., and Wu,
tural members are increasingly manufactured in minimills that T. (1995). "Toxic emissions indices for green design and inventory."
use almost exclusively steel scrap in electric arc furnaces to Envir. Sci. and Technol., 29(3), 86A-90A.
produce steel. If we assume that the resource inputs and the "Input-output accounts of the U.S. economy, 1987 Benchmark." (1994).
environmental outputs are lower from minimills than from in- U.S. Dept. of Commerce, Interindustry Economics Div., Compo Disk-
tegrated mills with basic oxygen furnaces, then the results of ettes, Washington, D.C.
Lave, L. B., Cobas, E., Hendrickson, C., and McMichael, F. C. (1995).
this study are skewed against steel. "Using input-output analysis to estimate economy-wide discharges."
A summary of environmental effects assessed in this article Envir. Sci. and Technol., 29(9), 420A-426A.
is given in Table 8. Many other environmental burdens have Lerner, A. C. (1996). "Infrastructure obsolescence and design service
not been included in this work due to the lack of data and, as life." J. Infrastruct. Sys., ASCE, 2(4), 153-161.

116/ JOURNAL OF INFRASTRUCTURE SYSTEMS / SEPTEMBER 1998

Downloaded 14 Jun 2012 to 128.237.190.83. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 Leontief, W. (1986). Input-output economics. Oxford University Press, U.S. Environmental Protection Agency (U.S. EPA), Solid Waste and
New York, N.Y. Emergency Response. (1993). "National biennial RCRA hazardous
Malin, N. (1997). "Steel or wood framing-Which way should we go?" waste report (based on 1989 data)." Rep. No. EPA 530-R-92-027,
Envir. Build. News, 3(4). Washington, D.C.
Marland, G., and Weinberg, A. M. (1988). "Longevity of infrastructure." U.S. Environmental Protection Agency (U.S. EPA). (1995). "1987-1993
Cities and their vital systems, J. H. Ausubel and R. Herman, eds., toxics release inventory." CD-ROM, EPA 749IC-95-004, Ofc. of Pol-
National Academy Press, Washington, D.C., 312-332. lution Prevention and Toxics, Washington, D.C.
"Means Building Construction Cost Data 1987." (1987). R. S. Means U.S. General Accounting Office (U.S. GAO). (1991). "Toxic chemicals
Co., Kingston, Mass. -EPA's toxic release inventory is useful but can be improved." GAOl
Portney, P. R. (1994). "The price is right: Making use of life cycle anal- RCED-91-12I, Washington, D.C.
yses." Issues in Sci. and Technol., 10(2), 69-75. Veshosky, D., and Nickerson, R. L. (1993). "Highway bridges: Life-cycle
Rainer, G. (1990). Understanding infrastructure. John Wiley & Sons, costs versus life-cycle performance." Better Roads, (May), 33-35.
Inc., New York, N.Y. Vigon, B. W., et al. (1993). "Life cycle assessment: Inventory guidelines
Spaans, L. (1997). "Bulb-tees make more efficient long-span bridges." and principles." Rep. No. EPA 60O/R-92/245, U.S. Envir. Protection
Civ. Engrg., ASCE, 47 -48. Agency, Washington, D.C.

JOURNAL OF INFRASTRUCTURE SYSTEMS / SEPTEMBER 1998 / 117

View publication stats Downloaded 14 Jun 2012 to 128.237.190.83. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org