United States
                Environmental Protection
                Agency
National Risk Management
Research Laboratory
Cincinnati, OH 45268
                Research and Development
EPA/600/SR-98/070    January 1999
&EPA    Project Summary

               A  Life-Cycle Impact Assessment
               Demonstration for the  GBU-24
               Duane Tolle, Bruce Vigon, and David Evers
                 The primary goal of this project was
               to develop and demonstrate a life-cycle
               impact assessment (LCIA) approach us-
               ing existing life-cycle inventory (LCI)
               data on one of the propellants, ener-
               getics, and pyrotechnic (PEP) materi-
               als of interest to the U.S. Department
               of Defense (DoD). Sponsorship for this
               study was from the Strategic Environ-
               mental Research and Development Pro-
               gram (SERDP) and involved coopera-
               tive efforts by  the DoD, U.S. Depart-
               ment of  Energy  (DOE),  and U.S.
               Environmental Protection Agency (EPA)
               in a program to develop technologies
               for clean production of PEP materials.
               Since the PEP program framework is
               strongly oriented around life-cycle as-
               sessment (LCA), a baseline LCI of the
               guided bomb unit-24  (GBU-24) made
               with  RDX explosives  had been con-
               ducted prior to this study and was se-
               lected as a test case for this LCIA.
                 An LCIA methodology and modeling
               approach was developed based on the
               Society of Environmental Toxicology
               and Chemistry's (SETAC's) Level  2/3
               equivalency assessment framework and
               applied to the  previously collected
               GBU-24 LCI data. The LCIA considered
               potential impacts on human health, eco-
               logical health, and resource depletion
               associated with  the GBU-24 life cycle.
               The approach included classification,
               characterization, normalization, and
               valuation. Quantitative equivalency fac-
               tors were  obtained  from the literature
               or developed for 11 of 14 potentially
               relevant impact categories. A regional
               scaling factor approach was developed
               to improve analysis of 4 of the 14 im-
               pact criteria, whose sensitivity to po-
               tential impacts varied on a regional ba-
               sis.
  The LCIA methodology based on im-
 pact equivalencies  described in this
 study provides a much more accurate
 approach to potential impact evaluation
 than  the "less-is-best"  approach
 (SETAC Level 1) using inventory data
 only. The method described in this re-
 port includes both regional scaling fac-
 tors to improve characterization accu-
 racy and geographically relevant nor-
 malization factors to provide perspec-
 tive. This bench-marking analysis can
 be  used for comparison with alterna-
 tive energetics.
  This project summary was  develop-
 ed by EPA's National Risk Management
 Research Laboratory to announce key
 findings of the research project that is
 fully documented in a separate report
 of the same title (see Project Report
 ordering information at back).
Introduction
  Development of future weapons sys-
tems will occur with consideration of envi-
ronmental  impacts during the acquisition
process. In fact, DoD policy has elevated
environmental considerations to a level of
importance equivalent to cost and perfor-
mance. In 1990, Congress established
SERDP as a  multi-agency effort to sup-
port environmental Research Design and
Development  (RD&D) programs. With
SERDP sponsorship, DoD, DOE, and EPA
have cooperated in a program to develop
technologies for the clean production  of
PEP materials. Along with the technology
oriented effort, a parallel activity has been
to develop and demonstrate analysis meth-
ods and tools  for estimating and manag-
ing the environmental  aspects of PEP
materials and the associated end items.
                                                               Printed on Recycled Paper

-------
  The framework for the activity has been
strongly oriented around  LCA. Thus,  a
baseline LCI of the current GBU-24 earth
penetrator  bomb was  conducted during
1993 and 1994 (the data basis was 1992
operations). That effort attempted to ad-
here very closely to the LCI methodology
described in SETAC and U.S. EPA tech-
nical  guideline publications.  Preliminary
results of that analysis have been reported
in several forums and publications.
  The purpose of this LCIA  demonstra-
tion  is to  develop and demonstrate the
LCIA methodology using the GBU-24 LCI
data collected previously. This is a baseline
or bench-marking analysis, which can be
used for future comparisons.
  The GBU-24 is  an earth penetrator
bomb equipped with  a laser guidance
package designed to  penetrate up to  6
feet of reinforced concrete. The BLU-109
bomb body is the largest physical compo-
nent and contributes the majority of the
material mass to the  system. The other
components listed were not included be-
cause they are minor in comparison  and
are readily reused  in  any event. Within
the BLU-109, the bomb case itself is the
largest source of material  (approximately
70% of the total weight) and efforts are
underway to evaluate ways to reduce pol-
lution from  its manufacture through recy-
cling  of the steel. Approximately 27% of
the total  comes from  the explosive fill.
The  PBXN-109 is a blend of four compo-
nents: CXM-7  explosive mix, aluminum
powder, thermoset plastic binder, and mis-
cellaneous  other blending and  forming
agents. About 3% of the mass is contrib-
uted by thermal insulation applied to the
bomb exterior and asphalt  interior liner.
  The GBU-24  life cycle  includes both
commercial as well as DoD facilities. Raw
materials are obtained for the energetic
materials production from commercial com-
modity chemical producers. The synthesis
of RDX,  together with the coating  and
blending needed to manufacture CXM-7,
is provided by Holston Army Ammunition
Plant (HSAAP) in Kingston, TN. The CXM-
7 is then shipped to McAlester Army Am-
munition Plant (MCAAP) in  McAlester, OK.
Load/assemble/pack (L/A/P)  operations,
which include blending the  CXM-7 with
aluminum and other additives to produce
the plastic-bonded explosive used for the
GBU-24,  are performed at MCAAP.  The
steel bomb bodies are also  shipped to
MCAAP from a commercial producer (Na-
tional Forge).
  Modeling of the GBU-specific manufac-
turing operations to obtain LCI data was
performed  in considerably greater detail
than for the commercial sector activities.
This was done for several reasons, not
the least of which was the fact that the
span of control of DoD for influencing such
major industrial activities as steel and am-
monia manufacture is limited.
  Once the bomb unit is manufactured it
undergoes qualification tests. Final assem-
bly of the  GBU-24 with fuse, guidance
control unit, adapter group, and  air-foil
group  is performed  on  aircraft carriers.
(This analysis assumed that the Navy ver-
sion of the GBU-24 is the system of inter-
est.)  Storage of the unit over the lifetime
of the weapon is included. Following re-
tirement, the item is decommissioned at
the Naval Service Warfare Center (NSWC)
using  waterjet extraction of  the fill  and
open  burning/detonation of the energetic
materials.
  The LCI/LCIA included activities from
cradle (raw  feedstock materials such as
ammonia)  to  grave (final  disposition
through disposal/recycling) for PEP end-
use items. The LCI data acquired included
primary information from government con-
trolled operations for the  manufacturing
and use operations and more generic in-
formation for ancillary operations. Ancil-
lary operations include feedstocks and ex-
ternal power grids.

Procedure
  The LCIA included an  initial scoping
process,  followed by classification, char-
acterization, normalization, and valuation.
Scoping included an evaluation of the data
available from  the LCI, a preliminary de-
termination  of the impacts  of  concern,
whether additional data are needed for
evaluating specific stressors, and a deci-
sion on the level(s) of impact analysis. In
order  to  facilitate the scoping,  stressor/
impact networks were prepared with pre-
liminary inventory data (including non-
quantitative) to determine the most appro-
priate impact categories for analysis and
to determine if the  LCI data are  in  the
correct form for impact analysis. The ba-
sis of comparison between two systems
in an LCA framework is the functional unit
(FU). The functional unit is determined by
the quantities associated with equivalent
performance levels of the alternatives. In
both the LCI and LCIA, the basis of the
analysis was one GBU-24 bomb.
  Classification  was  conducted after
scoping and is the process of linking or
assigning data from the LCI to individual
stressor categories within the three major
stressor categories of human health,  eco-
logical health, and resource depletion.
  Characterization involved the analysis
and estimation of the magnitude  of the
potential for stressors associated with the
baseline GBU-24 to contribute to each of
the impact  categories. The  equivalency
analysis approach functions by converting
a  large number of  individual stressors
within a  homogeneous  impact  category
into a single value,  by  comparing each
stressor with a reference material. The
procedure generally  involves multiplying
the appropriate equivalency factor by the
quantity of a resource or pollutant associ-
ated with a functional unit of GBU-24 (1
bomb) and summing  over all of the items
in  a classification category.
   For the Level 2 impact assessment (haz-
ard potential) evaluation used in this study,
a  limited subset  of the chemicals identi-
fied during the LCI had already been as-
signed impact equivalency  units in pub-
lished documents. Examples of groups of
chemicals that have  been  evaluated for
impact equivalency include nutrients, glo-
bal warming gases, ozone depletion gases,
acidification potential  chemicals, and pho-
tochemical oxidant precursors. Some of
the equivalency  factors  reported in the
literature were modified  by  application of
regional scaling factors.
   New impact equivalency (hazard poten-
tial) units for toxicity  impact criteria were
created  for some chemicals identified in
the baseline LCI by adapting the  hazard
ranking approach developed for the EPA
by the University of Tennessee to modify
the Level 3 Toxicity, Persistence, and Bio-
accumulation Potential Approach. This in-
cluded evaluation of impacts (e.g., toxicity
to  humans, fish, or wildlife) other than the
impacts evaluated in  Level  2, although a
few chemicals with multiple  impacts were
evaluated by both the Level 2 and 3 ap-
proaches.
   The carcinogenicity equivalency factor
is  based on the weight-of-evidence. (WOE)
for carcinogenicity as described  by either
the International Agency for Research on
Cancer (IARC) or the EPA. As suggested
in  the  University of  Tennessee  hazard
ranking report, a score of 1 to 5 was given
to  each  of the WOE groups.  Because
each agency has different ranking groups,
the equivalency score is based either on
an average of the scores for each agency
or the actual  score  if only one agency
ranked the chemical.
   Evaluation of the magnitude of resource
depletion impacts associated with the life-
cycle  of the  GBU-24 bomb started with
the resource  use  inventory  information
from the LCI.  Resources included in the
analysis involved  minerals  and  fossil fu-
els. These impacts were  evaluated from a
sustainability (time-metric  standpoint),
which considers the time to  exhaustion of
the resource. Information on the world re-
serve base  and production of  minerals
came from various documents by the U.S.
Geological Survey's  Minerals Information

-------
 Center  on the World  Wide Web. Infor-
 mation for energy sources came from the
 Annual Energy Review for 1994 by DOE's
 Energy  Information Administration. Water
 Use (consumption) was not selected as
 one of the primary impact categories, be-
 cause it was  known  at  the  outset that
 these data were not included in the inven-
 tory and because water availability is not
 considered to  be a problem at MCAAP,
 HSAAP, or NSWC.
   The equivalency factors for  the solid
 waste disposal impact criterion under land
 use are based on the  estimated volume
 calculated using the specific weight (in Ib/
 yd3) of each type of solid waste.  Since the
 LCI data for solid wastes are expressed
 as weight/functional units, multiplication of
 the weight and  inverse  of  the specific
 weight described the  landfill volume re-
 quired.
   Quantitative  equivalency factors were
 developed for 11  of the 14 impact catego-
 ries. A regional  scaling factor  approach
 was developed to improve analysis of 4 of
 the 14 impact criteria, whose sensitivity to
 potential impacts varies on a regional  ba-
 sis. Although the accuracy of the  impact
 scores for  these  four  impact  criteria is
 improved by this process, the  resulting
 impact scores are still not as accurate as
 the impact  scores for  the global criteria
 that are unaffected by regional differences
 in sensitivity. Since the impact  category
 for suspended particulates (PM10) included
 only  one stressor, the regional scaling
 analysis  was used without  a  need  for
 equivalency factors. The   inventory pro-
 vided by the model did not include data
 for emissions associated with the  ozone
 depletion impact criteria, even though the
 preliminary scoping analysis indicated that
 inventory data  for this impact  category
 should have been available. Land use as-
 sociated  with natural resource extraction
 was not evaluated due to  the difficulty in
 determining the quantity of land used  for
 many of the resources identified  in the
 inventory.
   Regional scaling factors were developed
 for the following four impact criteria: sus-
 pended paniculate (PM10) effects, acid
 deposition, smog creation, and eutrophi-
 cation. These impacts  have either a re-
 gional or local spatial resolution  because
 environmental conditions in different loca-
tions cause the  same emission quantity to
 have more  or  less impact. Some loca-
tions/regions may be highly sensitive to
one of these impacts and  other locations
may be only moderately affected or may
not experience  any impact at all  from the
same quantity of emissions. For each one
of these four impact categories,  different
levels of sensitivity throughout the U.S.
 were defined and linked with scaling fac-
 tors for use in  refining the final impact
 category scores.  In some cases these scal-
 ing factors were indicated on maps, based
 on  a composite of information,  such as
 sensitive receptors, emission sources, and
 emission deposition rates. In all four cases
 the scaling factors were averaged for each
 state according  to  the percent  of  area
 covered by all scaling factors for a given
 impact category within that state. These
 average state scaling factors were neces-
 sary for allocating emissions among states
 when specific facility  locations were not
 known or  too numerous (e.g., emissions
 associated with the national grid of elec-
 tric power generation plants).
   Normalization  is  recommended  after
 characterization and prior to valuation of
 LCIA data because aggregated sums per
 impact category need  to be expressed in
 equivalent terms before assigning valua-
 tion weight factors. The normalization step
 helps to put in  perspective the relative
 contribution  that a calculated  character-
 ization sum  for an indicator category
 makes relative to an actual environmental
 effect.
   The normalization approach involves the
 determination of factors that represent the
 total, annual, geographically relevant im-
 pact (expressed in Ibs/yr) for a given im-
 pact category. The goal is to develop sci-
 entifically defensible normalization factors,
 making use  of existing emissions or re-
 source extraction data. Impact categories
 are divided according to three spatial per-
 spectives:  global, regional,  or  local.  The
 global  impact categories  (e.g.,  global
 warming) are assumed to be independent
 of the geographic location in which emis-
 sions are released or  resources  are  ex-
 tracted. The regional  impact categories
 (e.g., acid rain) are relevant  to fairly large
 areas, but are clearly not global or limited
 to one site.  Thus,-data selected for the
 regional, normalization factors were based
 on the maximum annual state total impact
 (total emissions of relevant chemicals mul-
 tiplied by a regional scaling  factor). Local
 impact categories were limited to the three
 acute toxicity categories (e.g.,  terrestrial
 [wildlife] toxicity) because the area within
 which a single organism is  impacted  for
 each of these acute toxicity  categories is
 very small. The total impact used for deter-
 mining the  local normalization factor was
 considered to be the  maximum  annual
 emission of  relevant chemicals emitted
 from a single facility in the United States
 into  the environmental  medium of con-
 cern.
  Valuation involves assigning relative val-
 ues  or  weights  to  different impacts  so
they can be integrated across impact cat-
 egories for use  by decision  makers. It
 should be recognized that this is largely a
 subjective process, albeit one that is in-
 formed by knowledge of the nature of the
 issues involved.  The valuation  method
 used in this study is known as the Analyti-
 cal Hierarchy Process (AHP). AHP  is a
 recognized  methodology for  supporting
 decisions based  on relative preferences
 (importance) of pertinent factors. Environ-
 mental preferences are expressed by the
 valuation team in a pair-wise manner sup-
 ported by a software package known as
 Expert Choice™.


 Results

 Normalized Impact Criteria
 Scores
   Impact criteria scores (hazard potential)
 were developed for the baseline GBU-24
 production  process using the  inventory
 quantities of each stressor per functional
 unit. Impact criteria scores are calculated
 by multiplying the inventory quantity times
 the impact equivalency factor for each in-
 dividual chemical and then dividing by the
 total normalization factor for that impact
 category.
   The normalized impact scores per func-
 tional  unit in Table  1  indicate that the
 Terrestrial Toxicity impact category shows
 the greatest  normalized  impact score
 (4.26E-06) for the baseline GBU-24 pro-
 cess, when all impact categories are con-
 sidered to be  of equal importance  (i.e.,
 the  valuation weights have not been ap-
 plied). The  relative  contribution of each
 normalized impact score to the total nor-
 malized  impact score  for the  baseline
 GBU-24  process  is shown in the third
 column.  This  column  indicates that  the
 Carcinogenicity and  Terrestrial Toxicity
 impact categories contribute, respectively,
 41% and 42% of the total impact when all
 normalized impact scores  are considered
 of equal importance.

 Valuation-Weighted Impact
 Scores
  AHP valuation hierarchies were devel-
 oped to reflect two perspectives:  "policy"
 and "local". The "policy" perspective em-
 phasizes the global impacts of concern to
 a national policy maker. The "local"  per-
 spective emphasizes the local impacts of
 more concern to  someone siting a spe-
 cific facility. The valuation process as-
 signed weights to global,  regional,  and
 local, respectively, of 32%, 33%, and 35%
for the "policy" perspective, and 17%, 37%,
 and  47% for the "local" perspective. Final
weights were assigned for each of the 14
impact criteria.

-------
Table 1. Comparison of Normalized Impact
        Scores by Criteria for the  Baseline
        GBU-24 Production Process
Impact Category
               %of
  Normalized    Total
    Impact     For All
    Score     Scores
Ozone Depletion
Potential                NA^       0
Global Warming        2.55E-10      0
Resource Depletion     5.79E-09      0
Acid Rain             2.83E-08      0
Smog                2.28E-07      2
Suspended (PM10)
Particulates           1.79E-07      2
Human Inhalation
Toxtcity              2.84E-07      3
Carcinogenteity        4.21 E-06      41
Solid Waste Disposal
Land Use             1.14E-07      1
Resource Extraction/
Production Land Use       NA        0
Terrestrial
(Wildlife) Toxicity       4.26E-06      42
Aquatic (Fish) Toxlcity   7.06E-07      7
Eutrophteatton         1.64E-07      2
WNA « Data not avaitabte; relevant chemicals not listed
 in LCI.
  The weights developed by the AHP valu-
ation process were multiplied by the nor-
malized scores for each  impact category,
and these weighted, normalized  impact
scores were summed to  get a total score
for the baseline GBU-24 production pro-
cess. The  scores for each chemical or
resource contributing to  a particular im-
pact category were divided by the normal-
ization factor for that impact category. This
was  considered necessary before multi-
plying by the valuation weights to prevent
introduction of bias due to the large quan-
tities typically associated with resource ex-
traction  and use compared to the small
quantities typically associated with emis-
sions released after emission control de-
vices.
  The pie  diagrams shown in Figures 1
and 2 illustrate the percentages that each
weighted,  normalized impact  category
score contributes to the total weighted im-
pact score, respectively, for the "policy"
and  "local"  valuation perspectives. These
two figures show that carcinogenicity and
terrestrial toxicity are the top contributors
to the total potential impact of the baseline
GBU process, regardless of which of the
two valuation perspectives is used. In or-
der to reduce  the impact potential for the
baseline GBU  process, the individual
chemical emissions contributing the most
to these two  high scoring  impact cate-
gories are  logical choices to consider re-
ducing first.
  The LCIA methodology  based  on im-
pact equivalencies described in this report
provides a  much more accurate approach
to potential impact evaluation than the
"less-is-best" approach (SETAC Level 1)
using inventory data  only. The "less-is-
best" approach ignores the substantial dif-
ferences in impact potential between dif-
ferent chemicals contributing to the same
impact category.  For example, more hy-
droxide is released in wastewater per FU
than  ammonia,  but due to the higher
aquatic equivalency  factor for ammonia,
its normalized aquatic impact potential  is
greater.
  The "less-is-best"  approach  is also in-
accurate when entire impact  categories
are considered. If stressor quantities are
summed for  air emissions, water emis-
sions, solid wastes, and carcinogens, the
respective totals for each of these impact
categories  in Ibs  per FU are  2.69E+04,
3.54E-02, 1.27E+03, and  1.14E+01. This
Level 1 approach suggests that air emis-
sions associated  with the human health
inhalation toxicity  impact category have a
much greater impact than water emissions
associated with aquatic toxicity, or carci-
nogenic emissions associated  with carci-
nogenicity. However, valuation results for
both of the perspectives indicate that the
greatest potential  impact from these three
impact  categories is from carcinogenic
emissions.
  The method described in this report in-
cludes both regional  scaling factors to im-
prove characterization accuracy and geo-
graphically relevant normalization factors.
Although this method is expected to  be
somewhat less accurate than the generic
or site-specific exposure/effect assessment
approaches  using modeling,  it  requires
much less effort than either of these meth-
ods.
  The full  report  was submitted in fulfill-
ment of cooperative agreement CR822956
by Battelle under sponsorship  of the U.S.
Environmental Protection Agency.
             AQTOX7%
ACIDDEP 0%   EUTROPH 1%
           / SMOG 4%
                    SUSPPART 3%
                       INHLTOX 4%
   TERRTOX
   40%
                              AQTOX 7%
                                                                              EUTROPH1%
                                                                                       \  SMOG 3%
                                                     SUSPPART 3%
                                                         INHLTOX 4%
                          ^CARCINGN
                          '  41%
                    TERRTOX
                    40%
                    CARCINGN
                    42%
 Figure 1.  Impactcategorypercentagesoftotalimpactscoreweighted
          by the "Policy" perspective for the baseline GBU process.
                                             Figure 2.  ImpactCategorypercentagesoftotalimpactscoreweighted
                                                      by the "Local" perspective for the baseline GBU process.

-------

-------

-------

-------
  Duane Tolle, Bruce Vigon, and David Evers are with Battelle Columbus Labora-
  tories, Columbus, OH 43201-2693.
  Kenneth R. Stone is the EPA Project Officer (see below).
  The complete report, entitled "Life-Cycle Impact Assessment Demonstration for the
    GBU-24," (Order No. PS 99-102659; Cost $29.50 subject to change) will be
    available only from:
          National Technical Information Service
          5285 Port Royal Road
          Springfield, VA 22161
          Telephone: 703-487-4650
  The EPA Project Officer can be contacted at:
          National Risk Management Research Laboratory
          U.S. Environmental Protection Agency
          Cincinnati, OH 45268                  	
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
Official Business
Penalty for Private Use $300
     BULK RATE
POSTAGE & FEES PAID
         EPA
   PERMIT NO. G-35

EPA/600/SR-98/070

-------