Nuclear Waste
DOE Needs a Comprehensive Strategy and Guidance on Computer Models that Support Environmental Cleanup Decisions Gao ID: GAO-11-143 February 10, 2011The Department of Energy's (DOE) Office of Environmental Management (EM) is responsible for one of the world's largest cleanup programs: treatment and disposal of radioactive and hazardous waste created as a by-product of nuclear weapons production and energy research at sites across the country, such as EM's Hanford Site in Washington State and the Savannah River Site (SRS) in South Carolina. Computer models--which represent physical and biogeochemical processes as mathematical formulas--are one tool EM uses in the cleanups. GAO was asked to (1) describe how EM uses computer models in cleanup decisions; (2) evaluate how EM ensures the quality of its computer models; and (3) assess EM's overall strategy for managing its computer models. GAO analyzed the use of selected models in decisions at Hanford and SRS, reviewed numerous quality assurance documents, and interviewed DOE officials as well as contractors and regulators.
EM uses computer models to support key cleanup decisions. Because the results of these decisions can cost billions of dollars to implement and take decades to complete, it is crucial that the models are of the highest quality. Computer models provide critical information to EM's cleanup decision- making process, specifically to: (1) Analyze the potential effectiveness of cleanup alternatives. For example, computer models at SRS simulate the movement of contaminants through soil and groundwater and provide information used to predict the effectiveness of various cleanup strategies in reducing radioactive and hazardous material contamination. (2) Assess the likely performance of selected cleanup activities. After a particular cleanup strategy is selected, EM uses computer modeling to demonstrate that the selected strategy will be designed, constructed, and operated in a manner that protects workers, the public, and the environment. (3) Assist in planning and budgeting cleanups. EM also uses computer models to support lifecycle planning, scheduling, and budgeting for its cleanup activities. For example, a Hanford computer model simulates the retrieval and treatment of radioactive waste held in underground tanks and provides information used to project costs and schedules. EM uses general departmental policies and industry standards for ensuring quality, but they are not specific to computer models used in cleanup decisions. EM has not regularly performed periodic quality assurance assessments, as required by DOE policy, to oversee contractors' development and use of cleanup models and the models' associated software. In our review of eight cleanup decisions at Hanford and SRS that used computer modeling as a critical source of information, GAO found EM conducted required assessments of the quality of computer models in only three cases. In addition, citing flaws in a model EM uses to analyze soil and groundwater contamination, regulators from Washington state have told EM that it will no longer accept the use of this model for chemical exposure analysis at Hanford. EM does not have an overall strategy for managing its computer models. EM has recently begun some efforts to promote consistency in the use of models. For example, it is developing a set of state-of-the-art computer models to support soil and groundwater cleanup decisions across its sites. However, these efforts are still in early stages and are not part of a comprehensive, coordinated effort. Furthermore, although other federal agencies and DOE offices have recognized the importance of comprehensive guidance on the appropriate procedures for managing computer models, EM does not have such overarching guidance. As a result, EM may miss opportunities to improve the quality of computer models, reduce duplication between DOE sites, and share lessons learned across the nuclear weapons complex. GAO recommends that DOE (1) clarify specific quality assurance requirements for computer models used in environmental cleanup decision making; (2) ensure that the models are assessed for compliance with these requirements; and (3) develop a comprehensive strategy and guidance for managing its models. DOE agreed with our recommendations.
RecommendationsOur recommendations from this work are listed below with a Contact for more information. Status will change from "In process" to "Open," "Closed - implemented," or "Closed - not implemented" based on our follow up work.
Director: Eugene E. Aloise Team: Government Accountability Office: Natural Resources and Environment Phone: (202) 512-6870GAO-11-143, Nuclear Waste: DOE Needs a Comprehensive Strategy and Guidance on Computer Models that Support Environmental Cleanup Decisions
This is the accessible text file for GAO report number GAO-11-143
entitled 'Nuclear Waste: DOE Needs a Comprehensive Strategy and
Guidance on Computer Models that Support Environmental Cleanup
Decisions' which was released on March 11, 2011.
This text file was formatted by the U.S. Government Accountability
Office (GAO) to be accessible to users with visual impairments, as
part of a longer term project to improve GAO products' accessibility.
Every attempt has been made to maintain the structural and data
integrity of the original printed product. Accessibility features,
such as text descriptions of tables, consecutively numbered footnotes
placed at the end of the file, and the text of agency comment letters,
are provided but may not exactly duplicate the presentation or format
of the printed version. The portable document format (PDF) file is an
exact electronic replica of the printed version. We welcome your
feedback. Please E-mail your comments regarding the contents or
accessibility features of this document to Webmaster@gao.gov.
This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed
in its entirety without further permission from GAO. Because this work
may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this
material separately.
United States Government Accountability Office:
GAO:
Report to Congressional Requesters:
February 2011:
Nuclear Waste:
DOE Needs a Comprehensive Strategy and Guidance on Computer Models
that Support Environmental Cleanup Decisions:
GAO-11-143:
GAO Highlights:
Highlights of GAO-11-143, a report to congressional requesters.
Why GAO Did This Study:
The Department of Energy‘s (DOE) Office of Environmental Management
(EM) is responsible for one of the world‘s largest cleanup programs:
treatment and disposal of radioactive and hazardous waste created as a
by-product of nuclear weapons production and energy research at sites
across the country, such as EM‘s Hanford Site in Washington State and
the Savannah River Site (SRS) in South Carolina. Computer models”which
represent physical and biogeochemical processes as mathematical
formulas”are one tool EM uses in the cleanups. GAO was asked to (1)
describe how EM uses computer models in cleanup decisions; (2)
evaluate how EM ensures the quality of its computer models; and (3)
assess EM‘s overall strategy for managing its computer models. GAO
analyzed the use of selected models in decisions at Hanford and SRS,
reviewed numerous quality assurance documents, and interviewed DOE
officials as well as contractors and regulators.
What GAO Found:
EM uses computer models to support key cleanup decisions. Because the
results of these decisions can cost billions of dollars to implement
and take decades to complete, it is crucial that the models are of the
highest quality. Computer models provide critical information to EM‘s
cleanup decision-making process, specifically to:
* Analyze the potential effectiveness of cleanup alternatives. For
example, computer models at SRS simulate the movement of contaminants
through soil and groundwater and provide information used to predict
the effectiveness of various cleanup strategies in reducing
radioactive and hazardous material contamination.
* Assess the likely performance of selected cleanup activities. After
a particular cleanup strategy is selected, EM uses computer modeling
to demonstrate that the selected strategy will be designed,
constructed, and operated in a manner that protects workers, the
public, and the environment.
* Assist in planning and budgeting cleanups. EM also uses computer
models to support lifecycle planning, scheduling, and budgeting for
its cleanup activities. For example, a Hanford computer model
simulates the retrieval and treatment of radioactive waste held in
underground tanks and provides information used to project costs and
schedules.
EM uses general departmental policies and industry standards for
ensuring quality, but they are not specific to computer models used in
cleanup decisions. EM has not regularly performed periodic quality
assurance assessments, as required by DOE policy, to oversee
contractors‘ development and use of cleanup models and the models‘
associated software. In our review of eight cleanup decisions at
Hanford and SRS that used computer modeling as a critical source of
information, GAO found EM conducted required assessments of the
quality of computer models in only three cases. In addition, citing
flaws in a model EM uses to analyze soil and groundwater
contamination, regulators from Washington state have told EM that it
will no longer accept the use of this model for chemical exposure
analysis at Hanford.
EM does not have an overall strategy for managing its computer models.
EM has recently begun some efforts to promote consistency in the use
of models. For example, it is developing a set of state-of-the-art
computer models to support soil and groundwater cleanup decisions
across its sites. However, these efforts are still in early stages and
are not part of a comprehensive, coordinated effort. Furthermore,
although other federal agencies and DOE offices have recognized the
importance of comprehensive guidance on the appropriate procedures for
managing computer models, EM does not have such overarching guidance.
As a result, EM may miss opportunities to improve the quality of
computer models, reduce duplication between DOE sites, and share
lessons learned across the nuclear weapons complex.
What GAO Recommends:
GAO recommends that DOE (1) clarify specific quality assurance
requirements for computer models used in environmental cleanup
decision making; (2) ensure that the models are assessed for
compliance with these requirements; and (3) develop a comprehensive
strategy and guidance for managing its models. DOE agreed with our
recommendations.
View [hyperlink, http://www.gao.gov/products/GAO-11-143] or key
components. For more information, contact Gene Aloise at (202) 512-
3841 or aloisee@gao.gov.
[End of section]
Contents:
Letter:
Background:
Computer Models Provide Critical Information to EM's Environmental
Cleanup Decision-Making Process:
EM Has General Quality Policies for Its Computer Models, but It Has
Not Regularly Assessed Contractors' Implementation of Quality
Assurance Procedures:
EM Does Not Have an Overall Strategy and Guidance for Managing Its
Cleanup Models:
Conclusions:
Recommendations for Executive Action:
Agency Comments and Our Evaluation:
Appendix I: Scope and Methodology:
Appendix II: Functions of Key Models Used in Cleanup Decisions GAO
Reviewed at EM's Hanford and Savannah River Sites:
Appendix III: Comments from the Department of Energy:
Appendix IV: GAO Contact and Staff Acknowledgments:
Related GAO Products:
Abbreviations:
CERCLA: Comprehensive Environmental Response, Compensation, and
Liability Act of 1980, as amended:
DOE: Department of Energy:
EM: Office of Environmental Management:
EPA: Environmental Protection Agency:
NEPA: National Environmental Policy Act, as amended:
SRS: Savannah River Site:
[End of section]
United States Government Accountability Office:
Washington, DC 20548:
February 10, 2011:
The Honorable Fred Upton:
Chairman:
The Honorable Joe Barton:
Chairman Emeritus:
Committee on Energy and Commerce:
House of Representatives:
The Department of Energy's (DOE) Office of Environmental Management
(EM) is responsible for one of the world's largest environmental
cleanup programs, the treatment and disposal of radioactive and
hazardous waste created as a by-product of producing nuclear weapons
and energy research. The largest component of the cleanup mission is
the treatment and disposal of millions of gallons of highly
radioactive waste stored in aging and leak-prone underground tanks. In
addition, radioactive and hazardous contamination has migrated through
the soil into the groundwater, posing a significant threat to human
health and the environment. EM spends about $6 billion annually to
clean up its sites. As of February 2010, DOE estimated that the
overall cost to complete the entire cleanup mission will be between
$275 billion and $329 billion. Two DOE sites--the Hanford Site in
southeastern Washington state and the Savannah River Site (SRS) in
South Carolina--account for more than one-half of these annual costs
and about 60 percent of the total projected cost of the overall
cleanup of nuclear waste at DOE sites. As with nearly all of DOE's
missions, the majority of the work at these two sites is performed by
private firms under contract with DOE.
One tool EM uses to help decide how to clean up this radioactive and
hazardous waste is computer simulation modeling--hereafter referred to
as computer models--where the behavior of physical and biogeochemical
processes are described through the use of mathematical formulas. For
example, computer models may be used to simulate a process such as the
transport of contamination through the soil and groundwater or to
predict how long it will take to empty waste tanks in a certain
sequence. The results from these models often contribute to the basis
for cleanup decisions that can cost hundreds of millions of dollars to
implement.
The set of processes used to ensure the quality of computer software
and models--known as "quality assurance"--has been a concern in the
past. In 2000 and again in 2002, the Defense Nuclear Facilities Safety
Board raised concerns that DOE did not have adequate controls to
ensure the reliability of software used in nuclear facilities. The
Board noted that many systems used to maintain safety in nuclear or
hazardous facilities, such as ventilation system controls, rely on the
smooth operation of software to prevent accidents. Another concern
regarding software and modeling was raised at Hanford in 2006, when a
DOE headquarters review team found that the absence of quality
assurance oversight activities and the lack of formal data validation
and verification led to data inaccuracies in modeling used to support
the development of an environmental impact statement. These problems
prompted DOE to undertake a new modeling effort, delaying the
environmental impact statement.
In response to your request, this report (1) describes how EM uses
computer models in cleanup decisions; (2) evaluates how EM ensures the
quality of its computer models; and, (3) assesses EM's overall
strategy for managing its computer models. To address these
objectives, we gathered and reviewed information on the types of
cleanup decisions DOE has made at Hanford and SRS. For each site, we
selected examples of three types of decisions that were representative
of major decisions DOE has made at these sites between 2002 and 2010--
(1) decisions made under environmental statutes, including the
Comprehensive Environmental Response, Compensation, and Liability Act
of 1980, as amended (CERCLA)[Footnote 1]--which addresses specific
environmental remediation solutions for a cleanup site--and the
National Environmental Policy Act (NEPA), as amended[Footnote 2]--
under which DOE evaluates the impacts to human health and the
environment of proposed cleanup strategies and possible alternatives;
(2) performance assessments under DOE orders governing radioactive
waste management; and (3) budgeting and planning decisions for liquid
tank waste treatment and disposal. We then selected, based on input
from EM officials, the main models that were used to support these
decisions at the two sites. We obtained and reviewed documentation on
the computer models used and decisions made, and interviewed officials
from DOE headquarters to determine how the models were used in these
decisions. We analyzed this information to determine how the results
of computer models were used in making cleanup decisions and the
importance of the results. We also obtained and reviewed documentation
showing the standards the models were required to meet, as well as
DOE, contractor, and other quality assurance assessments indicating
whether these standards were met. We also interviewed officials from
the Environmental Protection Agency (EPA), the National Research
Council, the Defense Nuclear Facilities Safety Board, and other
organizations about existing standards for the use and implementation
of computer modeling software and modeling coordination strategies. We
visited both Hanford and SRS and spoke with EM officials and
contractor staff at both locations to better understand the use of
models in planning and cleanup decisions, as well as EM oversight of
the models. We focused our review on model standards and the use of
models in decisions, not on the quality of the models themselves or of
their output.
We conducted this performance audit from October 2009 to February
2011, in accordance with generally accepted government auditing
standards. Those standards require that we plan and perform the audit
to obtain sufficient, appropriate evidence to provide a reasonable
basis for our findings and conclusions based on our audit objectives.
We believe that the evidence obtained provides a reasonable basis for
our findings and conclusions based on our audit objectives.
Background:
Since the 1940s, DOE and its predecessors have operated a nationwide
complex of facilities used to research, design, and manufacture
nuclear weapons and related technologies. The environmental legacy of
nuclear weapons production at dozens of these sites across the United
States includes contaminated buildings, soil, water resources, and
large volumes of radioactive and hazardous wastes that require
treatment, storage, and disposal. The two sites that account for the
majority of the costs of the cleanup effort--Hanford and SRS--were
established in the 1940s and 1950s, respectively, to produce plutonium
and other nuclear materials needed to manufacture nuclear weapons. EM
manages cleanup projects at these and other sites that involve
multiple activities to treat and dispose of a wide variety of
radioactive and hazardous wastes.
Under federal and state laws, EM must clean up radioactive and
hazardous substances in accordance with specified standards and
regulatory requirements. EM carries out its cleanup activities under
the requirements of federal environmental laws that include, among
others, CERCLA and NEPA.[Footnote 3] CERCLA requires EM to evaluate
the nature and extent of contamination at the sites and determine what
cleanup remedies, if any, are necessary to protect human health and
the environment into the future. Under NEPA, EM must prepare an
environmental impact statement that assesses the environmental effects
for a proposed agency action, all reasonable alternatives, and the no-
action alternative. Under both the CERCLA and NEPA processes, EM
analyzes proposed remedial action alternatives according to
established criteria, invites and considers public comment, and
prepares a Record of Decision that documents the selected agency
action. If the cleanup method selected under CERCLA or NEPA will
result in disposal of waste at an on-site disposal facility, EM is
then required, under DOE's radioactive waste management order--DOE
Order 435.1--to ensure that waste management activities at each
disposal facility are designed, constructed, and operated in a manner
that protects workers, the public, and the environment.[Footnote 4] EM
does this by completing a "performance assessment" of the selected
cleanup method.[Footnote 5] To guide the implementation of selected
cleanup methods, EM and its contractors may prepare a "system plan"
that provides the basis for scheduling cleanup operations and
preparing budget requests. For example, both Hanford and SRS have
prepared system plans for treating and disposing of liquid radioactive
waste stored in aging and leak-prone underground tanks.
EM officials at DOE headquarters and field offices oversee cleanup
activities at the sites, but the work itself is carried out primarily
by private firms contracting with DOE. EM applies different approaches
to managing cleanup activities, depending on the type and extent of
contamination and state and/or federal requirements with which it
needs to comply. In addition, DOE has agreements with state and
federal regulators, known as Federal Facility Agreements, to clean up
the Hanford and SRS sites.[Footnote 6] The agreements lay out legally
binding milestones for completing major steps in the waste treatment
and cleanup process. EPA officials, as well as officials with
environmental agencies in the states where EM sites are located,
enforce applicable federal and state environmental laws and oversee
and advise EM on its cleanup efforts.
One tool EM uses in support of cleanup decision analyses is computer
modeling. Although the computer models used across EM sites vary, they
have certain common characteristics. In general, computer models are
based on mathematical formulas that are intended to reflect physical,
biogeochemical, mechanical, or thermal processes in simplified ways.
For example, a computer model can simulate the movement of
contamination through the soil and groundwater or simulate the
transfer of high-level radioactive waste from underground storage
tanks to facilities where the waste will be treated. Appendix II
details the key computer models used in the cleanup decisions we
reviewed at Hanford and SRS.
Computer Models Provide Critical Information to EM's Environmental
Cleanup Decision-Making Process:
EM uses computer models to provide critical information for its
decision-making process. First, computer models provide information
that EM uses to analyze the effectiveness of alternative actions to
clean up radioactive waste. Second, once a cleanup strategy has been
selected, computer modeling provides information that EM needs to
assess the performance of the selected cleanup strategy in reducing
risks to human health and the environment. Third, EM uses computer
models to simulate operations in the cleanup process, providing the
basis for planning cleanup efforts and for making annual budget
requests.
EM Uses Computer Models to Analyze the Potential Effectiveness of
Cleanup Alternatives:
EM's decision making for its cleanup efforts is based on meeting
federal and state requirements; input from state, local, and regional
stakeholders; and other considerations, including the costs of cleanup
actions. Computer models provide critical information that EM needs to
assess compliance with regulatory requirements when seeking to
identify and select alternatives for cleaning up radioactive and
hazardous wastes, as well as contaminated soil and groundwater at its
sites.
EM's cleanup decisions are guided by several federal and state
environmental laws, including CERCLA and NEPA, which both set forth
processes related to cleanup decisions. In the case of CERCLA, EM
determines the nature and extent of the contamination, assesses
various cleanup alternatives, and selects the best alternatives
according to evaluation criteria that include, among other things,
protection of human health and the environment, ease of implementing
the alternative, state and community acceptance, and cost. To
accomplish these steps, EM uses computer modeling to, among other
things, simulate the movement of contaminants through soil and
groundwater over many years assuming no cleanup action is taken.
Projected contamination levels, migration pathways, and contamination
travel timelines are provided by simulations and are evaluated to
determine whether regulatory standards will likely be exceeded in the
future. If action is needed, then modeling simulations may be
conducted for a number of different cleanup alternatives. For example,
EM used modeling to assess contamination and the potential
effectiveness of various cleanup strategies at SRS's C-Area
Burning/Rubble Pit. Used during the 1960s as a trash pit to dispose of
organic solvents, waste oils, paper, plastics, and rubble, SRS burned
the contents of the pit periodically to reduce its volume. Eventually,
SRS used the pit for the disposal of inert rubble, finally covering it
with two feet of soil in the early 1980s. However, the disposal of
these materials and periodic burning resulted in hazardous substance
contamination of the surrounding soil and groundwater. Between 1999
and 2004, EM implemented several actions to clean up the majority of
the area's contamination. Following these actions, EM used computer
models to simulate the movement of the remaining contamination through
the soil and groundwater over the next 1,000 years. Information
provided by this modeling helped EM to identify the remaining risks to
human health and the environment and to identify actions to clean up
the remaining contamination. Using this information, in conjunction
with other criteria such as additional site data, input from federal
and state regulators and the public, and the availability of an
appropriate cleanup technology, EM selected a final cleanup remedy.
This remedy, which is ongoing and combines several different cleanup
technologies, was estimated in 2008 to cost, in present-worth dollars,
about $1.9 million over a 70-year period.
In implementing CERCLA, DOE focuses on discrete facilities or areas
within a site that are being remediated, making limited assessments of
cumulative impacts. By contrast, under NEPA, EM generally prepares
environmental impact statements that assess the environmental impacts--
including cumulative impacts--of a proposed cleanup action, all
reasonable alternatives, and taking no action. For example, the
environmental impact statement for closing underground liquid
radioactive waste tanks at Hanford--which, as of November 2010 was
still in draft form--includes an analysis of the potential
environmental impact of various options for treating and disposing of
about 55 million gallons of mixed radioactive and hazardous waste and
closing 149 underground radioactive waste tanks. The draft
environmental impact statement includes an analysis of 11 tank waste
treatment and closure alternatives,[Footnote 7] including a no-action
alternative. These alternatives range in cost from about $3 billion to
nearly $252 billion, excluding the costs associated with the final
disposal of the treated waste.
In the draft environmental impact statement, EM used computer models
to simulate the movement of contamination through soil and groundwater
over a period of 10,000 years for each of the cleanup alternatives. As
with CERCLA modeling, the results of the computer models were used to
estimate the remaining risks to human health and the environment
following the completion of each cleanup alternative and these risks
were then compared with requirements. The results of these models will
be used along with other information such as input from regulators and
the public and the costs of each alternative when EM selects the
alternative it will eventually implement.
EM Uses Computer Models to Assess the Performance of Selected Cleanup
Activities:
After a particular cleanup alternative is selected, EM also uses
computer modeling to demonstrate that the cleanup activity will result
in reduced future contamination levels that meet regulatory
requirements. If the cleanup method selected under CERCLA or NEPA will
result in disposal of waste at an on-site disposal facility, EM is
then required, under DOE's radioactive waste management order--DOE
Order 435.1--to ensure that waste management activities at each
disposal facility are designed, constructed, and operated in a manner
that protects workers, the public, and the environment. To meet the
requirements of the order, EM completes a "performance
assessment"[Footnote 8] of the selected cleanup method. Under the
order, this performance assessment is to document that the disposal
facility is designed, constructed, and operated in a manner that
protects workers, the public, and the environment. The performance
assessment also is to project the release of contamination into the
soil and groundwater from a site after cleanup and must include
calculations of potential chemical doses to members of the public in
the future.
For example, in March 2010, SRS issued a performance assessment of a
cleanup and closure strategy for a group of 20 underground liquid
radioactive waste tanks, known as the F-Tank Farm.[Footnote 9] The
performance assessment evaluated closing the underground waste tanks
and filling them with a cement-like substance called grout--the
alternative selected following completion of SRS's 2002 environmental
impact statement. Computer modeling was used extensively to prepare
this performance assessment. Specifically, computer modeling was
performed using two different types of models. The first computer
model was used to perform human health and environmental risk
calculations and to calculate radiation doses that could be compared
to the maximum level allowed by federal and state requirements. The
second model was used to analyze sensitivities and uncertainties in
the results of the first model.
EM Uses Computer Modeling to Help Plan and Budget Cleanups:
EM also uses computer models for lifecycle planning, scheduling, and
budgeting for its cleanup activities. Computer models provide
important information that EM and its contractors use to develop
system plans that outline the schedules for cleanup activities at EM
sites. Outputs from computer models and databases are used to create
tables, charts, and schedules that are published in the system plans
and inform annual budget requests for cleanup activities.
For example, at Hanford, a computer model known as the Hanford Tank
Waste Operations Simulator is designed to track the retrieval and
treatment of over 55 million gallons of radioactive waste held in
underground storage tanks. According to the most recent Hanford tank
waste system plan, which was issued in November 2010, the model
projects the chemical and radiological characteristics of batches of
waste that are to be sent to a $12.2 billion waste treatment plant
that is being built at Hanford to treat this waste. The model also
provides scheduling information the contractor uses to project near-
and long-term costs and schedules.
Similarly, SRS uses a computer model known as SpaceMan Plus™ to
support the site's liquid waste system plan, which was issued in
January 2010.[Footnote 10] For example, project work schedules for
SRS's tank waste program are guided by this model. The model also
simulates how the tank farms integrate with waste processing
facilities and tracks the movement of waste throughout the liquid
waste system. Output from the model was used to provide tables and
schedules found in the appendixes of SRS's system plan that details
the specific cleanup activities that are to be accomplished. These
tables and schedules are used as part of the basis for determining the
costs of completing those activities. This information, in turn,
allows DOE and its contractors to generate annual budget requests.
EM Has General Quality Policies for Its Computer Models, but It Has
Not Regularly Assessed Contractors' Implementation of Quality
Assurance Procedures:
Although EM uses general departmental quality assurance policies and
standards that apply to computer models and relies on contractors to
implement specific procedures that reflect these policies and
standards, these policies and standards do not specifically provide
guidance on ensuring the quality of the computer models used in
cleanup decisions. Moreover, EM officials have not regularly performed
periodic quality assurance assessments, as required by DOE policy to
oversee contractors' development and use of cleanup models and the
models' associated software. In addition, DOE and others have
identified quality assurance problems. For example, the state of
Washington has cited flaws in a model EM uses to analyze soil and
groundwater contamination and has told EM that it will no longer
accept the use of this model for chemical exposure analysis at Hanford.
Although EM Has General Quality Assurance Standards, Its Oversight Is
Not Sufficient to Ensure the Quality of Cleanup Models:
DOE addresses quality through various departmental policies and
industry standards;[Footnote 11] however these policies and standards
do not specifically provide guidance on ensuring the quality of the
computer models used in cleanup decisions. Specifically, DOE's primary
quality assurance policy--DOE Order 414.1C[Footnote 12]--provides
general requirements EM and its contractors must meet to ensure all
work at the cleanup sites is carried out correctly and effectively,
including the development and use of computer models. These
requirements include developing a quality assurance program, training
staff how to check the quality of their work, and providing for
independent assessments of quality. A manual accompanying this order
describes acceptable, nonmandatory methods for specifically ensuring
quality of "safety software." Safety software is described in the
manual as software used to design, manage, or support nuclear
facilities. However, the manual is less clear on how to assure quality
in computer models. Furthermore, it does not clearly address the use
of computer software not considered as safety software, such as those
used by computer models that support DOE's cleanup decisions.
DOE's quality assurance order also requires contractors to select and
comply with an appropriate set of industry standards for all work,
including computer modeling. One common set of standards was developed
by the American Society of Mechanical Engineers and provides the
requirements necessary to ensure safety in nuclear facilities,
including the development and validation of computer models and
software that is used to design and operate such facilities.[Footnote
13] Initially, the American Society of Mechanical Engineers standards
were not mandatory for computer models and software used for cleanup
decisions, many of which are considered nonsafety software. These
standards were but one of many standards that contractors could choose
to use. However, as of November 2008, EM made the American Society of
Mechanical Engineers standards mandatory for all cleanup activities,
including modeling.
EM's contractors are to implement DOE's quality assurance requirements
using specific policies and procedures they develop. The specifics of
implementation vary from contractor to contractor. In the case of
computer software quality, a contractor is to include procedures for
testing and validating the software, ensuring changes to software are
properly documented, and correcting any errors. EM allows its
contractors to take a "graded approach" to quality procedures for
computer software, which means the contractor may adjust the rigor of
the quality procedures to match the importance of the software to
overall operations. According to documents we reviewed, computer
software that controls systems in a nuclear facility, for example,
would require more rigorous quality procedures than an administrative
payroll system, as any failure in the software controlling a nuclear
facility could result in potentially hazardous consequences to
workers, the public, and/or the environment.
EM is to oversee its contractors' implementation of quality standards
for computer models by performing periodic quality assurance
assessments, according to DOE's quality assurance order. These quality
assurance assessments are intended to ensure that computer models meet
DOE and accepted industry quality standards. In our review of eight
cleanup decisions at Hanford and SRS, we found EM had conducted only
three quality assurance assessments that addressed quality standards
for the models used in those decisions. For example, for three of the
four decisions we reviewed at SRS, DOE officials at SRS could not
provide quality assurance assessments that specifically addressed
whether the models used in those decision processes met DOE's quality
assurance requirements.[Footnote 14] DOE officials at SRS provided
three general quality assurance assessments, but these quality
assurance assessments did not specifically look at the cleanup models.
In contrast, the models for a March 2010 performance assessment
selecting a cleanup strategy to close underground liquid waste tanks
at SRS did receive a quality assurance assessment by a DOE
headquarters group established to review performance assessment
decisions.[Footnote 15] In particular, as part of the review, among
other things, the DOE group conducted a quality assurance assessment
that evaluated the quality of the computer models used in the
performance assessment and the degree to which the models complied
with DOE requirements and industry standards. A DOE quality assurance
official at SRS noted that the site relies primarily on its
contractors to perform quality assurance assessments of computer
models and their associated software.
Similarly, in our review of four cleanup decisions at Hanford, we
found that EM had performed assessments that addressed quality
standards for the models used in those decisions in only two cases. In
fact, one quality assurance assessment was only undertaken after a
contractor discovered data quality errors in 2005 in a computer model
used to support a prior environmental impact statement at Hanford.
According to a DOE quality assurance manager at Hanford, his office
conducts quality assurance assessments primarily on those computer
models and the associated software for which the failure would result
in significant safety consequences to workers, the public, and/or the
environment.
Some Reviews Have Revealed Quality Assurance Problems:
Concerns have been raised by DOE and others that EM does not have
complete assurance of the quality of the models. For example:
* Citing a number of flaws in a model DOE uses to analyze soil and
groundwater contamination at Hanford, the Washington state Department
of Ecology told DOE in February 2010 that it would no longer accept
the use of this model for chemical exposure analysis at Hanford. For
example, Ecology cited previous concerns that the model was not robust
enough to capture complexities of the movement of contamination
through the subsurface soil. We found that DOE had conducted no
specific quality assurance reviews on the model and its associated
software.
* EM headquarters officials conducted two technical reviews in 2009 of
planning models used for tank waste operations at Hanford and SRS.
[Footnote 16] The review of the Hanford planning model found that the
model has limited ability to sufficiently predict the composition of
the contaminated waste as it is prepared for the treatment processes.
The review team cautioned that this limitation raised a significant
risk that, when actual waste treatment operations started at the site,
the waste may not meet the acceptance requirements for processing by
Hanford's treatment facility. In addition, the review of SRS's
planning model found that, although the data the model provided on
tank waste operations were reasonable, the model did not have the
ability to optimize operating scenarios, which hampered the site's
long-term planning abilities.
* A March 2010 independent review commissioned by a Hanford citizen's
group raised concerns about a model used in the preparation of a draft
environmental impact statement of alternatives for closing Hanford's
waste tanks.[Footnote 17] These concerns, based on reviewing the draft
statement, included insufficient documentation of the quality
assurance processes followed for the model and that modeling
uncertainties were inadequately quantified. The review concluded that
the environmental impact statement was insufficiently precise to be
used to make a cleanup decision.
Where DOE has conducted quality assurance assessments, it has found
that contractors did not always implement quality requirements
consistently. Furthermore, in their own internal reviews, contractors
have noted problems with the implementation of quality assurance
requirements. Problems noted in DOE's and contractors' quality
assurance assessments include:
* Inadequate documentation. A 2007 software quality review conducted
by DOE at Hanford found implementation problems, including inadequate
documentation and improper training for personnel in quality
procedures. At SRS, two general software quality assurance reviews
performed by DOE in 2004 found that while contractors generally met
quality requirements, documentation was sometimes lacking or
improperly prepared. A similar 2007 DOE review at SRS found a good
software quality program overall, but listed a number of deficiencies
including inadequate software plans and procedures.
* Not following correct procedures. A 2007 DOE review of a Hanford
contractor's software quality assurance program found, among other
things, that not all contractor personnel fully understood software
quality requirements. The report stated that, although software
quality assurance training had been provided, personnel did not follow
procedures in managing, maintaining, and overseeing software quality.
For example, the report cited an example of a spreadsheet in which
data input cells were not properly locked, in violation of procedures.
In addition, the report noted that software documentation was not
periodically updated, as required, because staff did not fully
understand the procedures.
* Incorrect quality assurance grading. In some cases, contractors did
not always correctly determine the level of rigor needed to ensure the
quality of computer models and their associated software. For example,
a 2007 internal contractor review at Hanford found that 23 of 138
software codes registered in a central repository were incorrectly
designated as nonsafety software, when in fact they should have been
considered safety software. As a result, the quality assurance
procedures appropriate for a given level of risk may not have always
been applied.
EM Does Not Have an Overall Strategy and Guidance for Managing Its
Cleanup Models:
Although EM has recently begun some efforts to promote consistency in
the use of models across its various sites, these efforts are still in
early stages and, to date, some have had limited involvement of
modeling officials at the sites and federal, state, and local
stakeholders who are affected by decisions made using the output of
computer models. In addition, these efforts are not part of a
comprehensive, coordinated effort to improve the management of
computer models across EM. In the absence of such a strategy, EM also
does not have overarching guidance promoting consistency in modeling
management, development, and use across EM's sites.
EM Has Some Initiatives to Improve Management of Its Cleanup Models,
but They Are Not Part of a Comprehensive, Coordinated Strategy:
EM has begun some efforts to improve the use of computer models across
its various sites. For example, EM, in fiscal year 2010, began
developing a set of state-of-the-art computer models to support soil
and groundwater cleanup across the nuclear weapons complex. According
to EM officials and documentation they provided, this initiative,
called the Advanced Simulation Capability for Environmental
Management, will allow EM to provide more sophisticated analysis of
soil and groundwater contamination for cleanup decisions. Although the
initiative's director told us that the goal is to encourage all sites
to use these models for all of their soil and groundwater analysis, he
noted that there are no plans to make using these models mandatory.
Moreover, SRS has created a forum for improving consistency in
groundwater computer modeling performed at the site. According to the
charter document, the forum, called the Groundwater Modeling
Consistency Team, was formed in 2006 following the discovery of
inconsistencies in the data used in groundwater computer modeling
conducted at Hanford in support of the preparation of an environmental
impact statement under NEPA. The group, which is made up of DOE and
contractor officials, reviews software codes, model inputs, and model
assumptions to promote sitewide consistency in the management of
computer models.
Although these efforts may help improve EM's use of computer models,
they are largely still in early stages. In addition, according to EM
officials, some of these efforts have, to date, had limited
involvement of modeling officials at EM's sites and of federal, state,
and local stakeholders who are affected by decisions made using the
output of computer models. Furthermore, they are not part of a
comprehensive, coordinated effort to improve the consistency of
computer models and reduce duplication across EM's various sites. For
example, we found that different models are used to perform similar
functions not only between EM sites, but also within sites. At SRS,
one contractor uses a set of models to perform soil and groundwater
analyses when evaluating the potential effectiveness of cleanup
alternatives under CERCLA and NEPA, while another contractor uses a
different set of models to perform similar analyses for performance
assessments under DOE's radioactive waste management order. Each
contractor has its own set of procedures for developing and using each
computer model. Officials from both contractors told us that they use
different models because state and federal regulators have only
approved the use of certain models for specific types of cleanup
decisions. Issues with consistency and duplication of effort in the
use of computer models have also been noted by others. For example, a
February 2010 DOE review noted that five major DOE sites use 28
different models to analyze groundwater and subsurface contamination
when preparing performance assessments under DOE's radioactive waste
management order. DOE officials told us that past modeling practices
have resulted in conflicting assumptions and data sets, as well as
different approaches to uncertainty analyses. In addition, a September
2009 DOE technical review of the Hanford tank waste modeling system
raised concerns that two models at Hanford that share data use
different assumptions that could lead to inconsistencies between the
two. As a result, the Hanford waste treatment system plan, which is
based on the output of one of these models, may not reflect the most
current information.
In contrast, other federal agencies and DOE offices have taken steps
to improve consistency and reduce duplication as part of a
comprehensive, coordinated strategy to manage the use of computer
models. For example, EPA organized a Center for Regulatory
Environmental Modeling in 2000 as part of a centralized effort to
bring consistency to model development, evaluation, and usage across
the agency. The Center brings together senior managers, modelers, and
scientists from across the agency to address modeling issues. Among
its tasks are to help the agency (1) establish and implement criteria
so that model-based decisions satisfy regulatory requirements; (2)
implement best management practices to use models consistently and
appropriately; (3) facilitate information exchange among model
developers and users so models can be continuously improved; and (4)
prepare for the next generation of environmental models. According to
a DOE official, EM does not have a central coordination point similar
to EPA's.
Within DOE, the Office of Nuclear Energy recently established an
initiative--the Nuclear Energy Modeling and Simulation Energy
Innovation Hub--that provides a centralized forum for nuclear energy
modelers. According to the director of the Office of Nuclear Energy's
Office of Advanced Modeling and Simulation, the hub will provide a
more centrally coordinated effort to bring together modeling and
simulation expertise to address issues associated with the next
generation of nuclear reactors. Similar comprehensive, coordinated
efforts are lacking within EM and, as a result, EM may be losing
opportunities to improve the quality of its models, reduce
duplication, keep abreast of emerging computer modeling and cleanup
technologies, and share lessons learned across EM's sites.
Other Federal Agencies and DOE Offices Have Recognized the Need for
Comprehensive Modeling Guidance:
The need for specific guidance for ensuring the careful management of
computer models used in decision making is not new. As early as 1976,
we reported on the government's use of computer models and found that
the lack of guidance contributed to ineffective and inefficient use of
computer models.[Footnote 18] We noted that guidance should define the
problem to be solved, specify the assumptions and limitations of the
model, and provide methods to test whether the model reasonably
describes the physical system it is modeling.
More recently, a 2007 National Research Council study of modeling at
EPA laid out guidelines to improve environmental regulatory computer
modeling.[Footnote 19] The study noted that adoption of a
comprehensive strategy for evaluating and refining EPA's models could
help the agency add credibility to decisions based on modeling
results. It also noted several key principles to follow for model
development, evaluation, and selection. Moreover, the study
recommended that peer review be considered as an important tool for
improving model quality. According to the study, a peer review should
entail not only an evaluation of the model and its output, but also a
review of the model's origin and its history. The study also made
recommendations on quantifying and communicating uncertainty in model
results to better communicate a model's limitations to stakeholders
affected by decisions made using the results of computer models.
EPA has taken action to develop specific guidance, issuing a guide in
2009 addressing the management, development, and use of computer
modeling used in making environmental regulatory decisions.[Footnote
20] In this guidance, EPA developed a set of recommended best
practices to help modelers effectively use computer models. The
guidance defines the role of computer models in the public policy
process, discusses appropriate ways of dealing with uncertainty,
establishes criteria for peer review, and addresses quality assurance
procedures for computer modeling.
Even within DOE, another office outside of EM has recognized the need
for specific guidance for managing computer models. Specifically,
DOE's Office of Civilian Radioactive Waste Management specified in its
quality assurance requirements several requirements for computer
models.[Footnote 21] These requirements included clearly defining the
model's objective, documenting alternative models that could be used
and the rationales for not using them, and discussed a model's
limitations and uncertainties. In addition, the office specified in
its requirements that, among other things, a computer model receive a
technical review through a peer review or publication in a
professional journal.
Although the importance of comprehensive guidelines for managing
computer models is well established, according to its officials, EM
does not have such overarching guidance. As previously discussed, EM
does have a manual accompanying its quality assurance order that
describes acceptable methods for specifically ensuring the quality of
safety software. However, the manual does not generally address models
used in cleanup decisions. EM also has guidance addressing the
management of computer models used in conducting performance
assessments under its radioactive waste management order.
Specifically, a DOE headquarters group that is charged with reviewing
decisions made under this order--the Low-Level Waste Disposal Facility
Federal Review Group--has developed a manual that contains guidance
on, for example, ensuring that input data to computer models are
described and are traceable to sources derived from, among other
things, field data from the site and referenced literature that is
applicable to the site. However, this guidance does not apply to
computer models used to analyze the potential effectiveness of cleanup
alternatives under CERCLA or NEPA or to computer models used for
planning, scheduling, and budgeting purposes. As a result, computer
models developed at various DOE sites do not have consistent criteria
to define the role of the model in the decision-making process,
consistent ways of dealing with uncertainties and a model's
limitations, and mechanisms to ensure computer model quality, such as
quality assurance assessments and peer review.
Conclusions:
EM's computer models provide critical information that is needed to
make significant decisions about how to clean up the radioactive and
hazardous legacy waste across the country. However, EM's oversight of
the quality of these models and its management of the development,
evaluation, and use of the models has not always been commensurate
with the models' importance. Because the decisions EM makes must
protect human health and the environment for thousands of years into
the future, it is critical that the models on which EM bases its
decisions are of the highest quality possible. In addition, because
these cleanup efforts will take decades and cost billions of dollars,
it is also important that models used for planning, scheduling, and
budgeting purposes provide the most accurate data possible for EM and
Congress to make informed decisions on cleanup activities.
EM's failure to fully oversee its contractors' implementation of
quality assurance procedures has led to a reduced level of confidence
that the models reasonably represent the conditions they are meant to
simulate. In several cases, we found necessary quality assurance
reviews were not conducted. In others, reviews found that quality
assurance procedures were inadequately implemented. Because existing
quality assurance requirements that are applied to EM's computer
models have not been adequately implemented and, in some cases, are
insufficiently understood by its contractors, EM and its contractors
do not have an effective mechanism to provide the public and other EM
stakeholders with assurance of a model's quality.
To its credit, EM is beginning to undertake efforts to improve the
consistency of models across the nuclear weapons complex. However,
some of these efforts are still in their infancy, and it remains to be
seen whether any improvements in EM's management of its models will
result. We recognize that every site has its unique conditions and
challenges and that a one-size-fits-all approach to modeling would not
be appropriate. Nevertheless, there is room for additional consistency
in model development and implementation, as well as a mechanism for
sharing lessons learned among DOE's various sites. For a number of
years, other federal agencies and offices within DOE have recognized
the importance of a comprehensive guidance for managing computer
models. Without a comprehensive strategy and modeling guidance, EM may
miss opportunities to improve the quality of computer models, promote
consistency, reduce duplication across DOE sites, and share lessons
learned.
Recommendations for Executive Action:
To help EM increase confidence in the quality of information provided
to the public and its stakeholders resulting from the use of computer
modeling, we recommend the Secretary of Energy take the following
three actions:
* Clarify specific quality assurance requirements for computer models
used in environmental cleanup decisions, including to analyze the
potential effectiveness of cleanup alternatives, assess the
performance of selected cleanup activities, and assist in planning and
budgeting cleanup activities.
* Ensure that the models are assessed for compliance with these
requirements.
* Develop a comprehensive strategy and guidance for the management of
computer models to promote consistency, reduce duplication, and ensure
sharing of lessons learned.
Agency Comments and Our Evaluation:
We provided a draft of this report to DOE for its review and comment.
In its written comments, DOE agreed with our recommendations and
stated that modeling is an important component of management analysis
and decision making for the department. DOE noted that it is committed
to continuous improvement in model development and application and
commented that our recommendations will strengthen its modeling
efforts.
DOE stated in its comments that it disagreed with the draft report's
assertion that its directives and standards fall short for the
development and management of computer models. DOE commented that its
quality assurance directives apply directly to the development,
coding, and validation of safety and nonsafety computer models used in
cleanup decisions and that EM has interpreted and applied these
directives and accompanying standards to develop its quality program.
We agree with DOE, and our draft report noted, that DOE addresses
quality through various departmental policies and industry standards.
However, these directives do not provide specific guidance to EM on
assuring quality of the cleanup models themselves, guidance that other
agencies and offices within DOE have developed. In particular, DOE's
primary quality assurance policy--DOE Order 414.1C--addresses general
standards that EM and its contractors must meet to ensure all work at
its sites is carried out effectively, but is vague on the specific
steps that must be followed to ensure the quality of models used in
cleanup decisions. In addition, as our draft report noted, a manual
accompanying this order describes acceptable, nonmandatory methods for
specifically ensuring quality of safety software. However, the manual
is less clear on the use of computer software not considered as safety
software, such as those used by computer models that support DOE's
cleanup decisions. Our recommendation that DOE clarify the specific
quality assurance requirements for computer models used in
environmental cleanup decisions is intended to address these problems.
DOE's comments also provided additional information on the
department's oversight of computer models, initiatives it is
undertaking to improve its modeling efforts, and the specific steps it
plans to take to address our recommendations. DOE also provided
technical comments that we incorporated in the report as appropriate.
DOE's written comments are presented in appendix III.
As agreed with your office, unless you publicly announce the contents
of this report earlier, we plan no further distribution until 30 days
from the report date. At that time, we will send copies of this report
to the appropriate congressional committees; the Secretary of Energy;
the Director, Office of Management and Budget; and other interested
parties. In addition, the report will be available at no charge on the
GAO Web site at [hyperlink, http://www.gao.gov].
If you or your staffs have any questions regarding this report, please
contact me at (202) 512-3841 or aloisee@gao.gov. Contact points for
our Offices of Congressional Relations and Public Affairs may be found
on the last page of this report. Key contributors to this report are
listed in appendix IV.
Signed by:
Gene Aloise:
Director, Natural Resources and Environment:
[End of section]
Appendix I: Scope and Methodology:
To determine how the Department of Energy's (DOE) Office of
Environmental Management (EM) uses computer modeling in cleanup
decisions, we focused on cleanup decisions EM has made at its Hanford
Site in Washington state and Savannah River Site (SRS) in South
Carolina because together these two sites account for more than one-
half of EM's annual cleanup spending and approximately 60 percent of
the total estimated cost of approximately $275 billion to $329 billion
to clean up the entire nuclear weapons complex. We focused our review
on decisions made in two major areas that represent the largest and
most significant elements of the cleanup program at these two sites.
The first is cleanup of radioactive and hazardous waste stored in
underground tanks, which DOE has determined poses the most significant
environmental safety and health threat in the cleanup program. DOE
estimates cleaning up tank waste at the sites will cost between $87
billion and $117 billion, making it the largest cost element of EM's
cleanup program. Second, both sites have significant contamination to
soil and groundwater, which DOE estimates will cost more than $12
billion to remediate. For each site, we selected three types of
decisions that were representative of major decisions made at these
sites between 2002 and 2010--(1) decisions made under environmental
statutes, including the Comprehensive Environmental Response,
Compensation, and Liability Act of 1980, as amended (CERCLA)--which
addresses specific environmental remediation solutions for a cleanup
site--and the National Environmental Policy Act, as amended (NEPA)--
under which DOE evaluates the impacts to human health and the
environment of proposed cleanup strategies and possible alternatives;
(2) performance assessments under DOE orders governing radioactive
waste management; and (3) cleanup budgeting and planning decisions. We
reviewed publicly available information from regulators and
interviewed DOE officials and contractor staff to identify the most
recent decisions for each of the three types of decisions selected for
review at each site. We reviewed these decisions to identify the most
recent decision that included the use of computer modeling. We then
selected, based on input by EM officials, the main models used to
support these decisions at the two sites. We visited both Hanford and
SRS and spoke with both EM officials and contractor staff there to
better understand the use of models in planning and cleanup decisions
and DOE's oversight of the models. We obtained demonstrations of these
models, as well as information on how they were used in decision
making. We obtained and reviewed the decision documents, as well as
modeling studies, notes of meetings between DOE and its regulators to
develop models, and other documentation showing how the models were
used in decisions. We interviewed officials from DOE headquarters and
the two sites, as well as contractor staff, to determine how the
models work and how they were used in these decisions. We analyzed
this information to determine how the results of computer models were
used in making cleanup decisions, the importance of modeling in the
selection of a cleanup strategy, and other factors that contributed to
the selection of a cleanup strategy.
To evaluate how EM determines the quality of the computer models used
in cleanup decision making, we obtained and reviewed documentation
showing the standards the models were required to meet. We gathered
documentation on DOE standards, as well as policies and procedures
from contractors overseeing the models. We discussed computer model
and software standards with EM officials from EM's sites, contractors
at the sites, and headquarters officials. We also interviewed
officials from the Defense Nuclear Facilities Safety Board, the
National Research Council, the Environmental Protection Agency, and
the Washington state Department of Ecology about existing standards
for the use and implementation of computer modeling and its associated
software. We analyzed EM policies and contractor procedures to
determine what quality assurance standards exist to address the
quality of computer models. We also requested from EM and its
contractors all assessments that were conducted on computer models
used in the decisions we were reviewing, indicating whether quality
standards were met. In general, the assessments we reviewed were
largely conducted by the contractors, regulators, or external sources,
such as consultants. These reviews ranged from contractor-performed
assessments of the implementation of quality standards for software,
to federal and state regulator comments on the modeling output used to
develop alternatives in a regulatory package, to an outside consultant-
performed review on the appropriateness of modeling for selecting a
preferred alternative from an environmental impact statement prepared
under NEPA. We analyzed these assessments to understand the level of
oversight EM provided to assure model and software quality, as well as
the extent to which contractors were implementing quality procedures.
To address EM's overall strategy for managing computer models that are
used in cleanup decisions, we interviewed DOE officials from
headquarters and from each site. We also interviewed officials from
the Environmental Protection Agency, National Research Council, DOE's
Office of Nuclear Energy, and DOE's Office of Civilian Radioactive
Waste Management about the implementation of computer modeling
guidance and modeling coordination strategies. We reviewed modeling
guidance from these organizations, as well as from the Office of
Management and Budget. We focused our review on model quality
assurance standards and the use of models in decision making, not on
the quality of the models themselves or of their output.
We conducted this performance audit from October 2009 to February
2011, in accordance with generally accepted government auditing
standards. Those standards require that we plan and perform the audit
to obtain sufficient, appropriate evidence to provide a reasonable
basis for our findings and conclusions based on our audit objectives.
We believe that the evidence obtained provides a reasonable basis for
our findings and conclusions based on our audit objectives.
[End of section]
Appendix II: Functions of Key Models Used in Cleanup Decisions GAO
Reviewed at EM's Hanford and Savannah River Sites:
Type of cleanup decision: Record of Decision under Comprehensive
Environmental Response, Compensation, and Liability Act of 1980, as
amended (CERCLA);
Models used at Hanford: RESRAD;
Description of how Hanford uses the model in the cleanup decision:
Uses one-dimensional, simplified model of contaminant transport from
the contaminated zone, through the vadose zone, to the aquifer;
Models used at the Savannah River Site: MODFLOW;
Description of how the Savannah River Site uses the model in the
cleanup decision: Approximates groundwater flow in a three-dimensional
grid. Used to estimate groundwater concentrations for contaminants
over time.
Type of cleanup decision: Record of Decision under Comprehensive
Environmental Response, Compensation, and Liability Act of 1980, as
amended (CERCLA);
Models used at Hanford: STOMP;
Description of how Hanford uses the model in the cleanup decision:
Used with RESRAD in performing contaminant transport-to-groundwater
evaluations;
Models used at the Savannah River Site: SEASOIL;
Description of how the Savannah River Site uses the model in the
cleanup decision: Simulates vertical transport of contaminants from
source, through the vadose zone, to the water table aquifer.
Type of cleanup decision: Environmental Impact Statement/Record of
Decision under National Environmental Policy Act (NEPA);
Models used at Hanford: MODFLOW;
Description of how Hanford uses the model in the cleanup decision:
Simulates the groundwater flow field in three dimensions--two
horizontal and one vertical--and contaminant transport from points of
contact with groundwater at various times to various locations;
Models used at the Savannah River Site: MEPAS;
Description of how the Savannah River Site uses the model in the
cleanup decision: Simulates fluid flow and contaminant transport in a
three-dimensional grid in the vadose zone and the saturated zone.
Transport results used to calculate groundwater concentrations for
multiple contaminants over time.
Type of cleanup decision: Environmental Impact Statement/Record of
Decision under National Environmental Policy Act (NEPA);
Models used at Hanford: STOMP;
Description of how Hanford uses the model in the cleanup decision:
Type of cleanup decision: Simulates three-dimensional, nonlinear water
and contaminant transport through the vadose zone over time;
Description of how the Savannah River Site uses the model in the
cleanup decision: Simulates fluid flow and contaminant transport in a
three-dimensional grid in the vadose zone and the saturated zone.
Transport results used to calculate groundwater concentrations for
multiple contaminants over time.
Type of cleanup decision: Environmental Impact Statement/Record of
Decision under National Environmental Policy Act (NEPA);
Models used at Hanford: HTWOS;
Description of how Hanford uses the model in the cleanup decision:
Provided assumptions that were used in the Hanford Environmental
Impact Statement as the basis for the number and location of waste
receiver facilities;
Description of how the Savannah River Site uses the model in the
cleanup decision: Simulates fluid flow and contaminant transport in a
three-dimensional grid in the vadose zone and the saturated zone.
Transport results used to calculate groundwater concentrations for
multiple contaminants over time.
Type of cleanup decision: Performance Assessment under DOE's
Radioactive Waste Management Order--DOE Order 435.1;
Models used at Hanford: DMT;
Description of how Hanford uses the model in the cleanup decision: A
graphical interface model that uses STOMP modeling output to
graphically display risk results. Used when calculating groundwater
concentrations of selected contaminants, predicting risk, and
comparing to regulatory criteria;
Models used at the Savannah River Site: PORFLOW;
Description of how the Savannah River Site uses the model in the
cleanup decision: Used to calculate radiological doses and perform
radiological and human health and ecological risk evaluation.
Type of cleanup decision: Performance Assessment under DOE's
Radioactive Waste Management Order--DOE Order 435.1;
Models used at Hanford: STOMP;
Description of how Hanford uses the model in the cleanup decision:
Type of cleanup decision: Modeled flow and transport of contaminants
through the vadose zone and groundwater; Provided inventory estimates
at tank closures for tank residue, as well as the concentration of
radionuclides and hazardous chemicals in tank retrieval solutions;
Models used at the Savannah River Site: GoldSim;
Description of how the Savannah River Site uses the model in the
cleanup decision: Used with PORFLOW to assist in developing
uncertainty and sensitivity analysis. Also used to calculate
radiological doses using either concentration results from PORFLOW or
GoldSim.
Type of cleanup decision: Planning and budgeting;
Models used at Hanford: HTWOS;
Description of how Hanford uses the model in the cleanup decision:
Simulates the movement of contaminated waste stored in underground
tanks as it is retrieved, prepared for treatment, and processed
through Hanford's under-construction waste treatment plant;
Models used at the Savannah River Site: SpaceMan Plus™;
Description of how the Savannah River Site uses the model in the
cleanup decision: Simulates the operation of the process in the liquid
tank waste system, from waste retrieval to waste processing, through
the site's waste processing facilities.
Source: GAO analysis of information from DOE.
[End of table]
[End of section]
Appendix III: Comments from the Department of Energy:
Department of Energy:
Washington, DC 20585:
January 19, 2011:
Mr. Gene Aloise:
Director, Natural Resources and Environment:
U.S. Government Accountability Office:
Washington, DC 20548:
Dear Mr. Aloise:
Thank you for the opportunity to review and comment on the draft
report on the Department of Energy's (DOE) Office of Environmental
Management (EM) modeling program, "Nuclear Waste: DOE Needs a
Comprehensive Strategy and Guidance on Computer Models that Support
Environmental Cleanup Decisions." Modeling is an important component
of management analysis and decision making for the Department's highly
complex and varied cleanup activities. As such, EM is committed to a
process of continuous improvement in both model development and
application by both Federal and contractor employees.
The draft U.S. Government Accountability Office (GAO) report
recognizes EM's adherence to the Department's directives and
accompanying industry standards for the development of its corporate
quality program. However, we disagree with the report's assertion that
these directives and standards fall short for the development and
management of computer models. DOE quality assurance directives
(DOE 0 414.1C, 10 CFR 830, and ASME NQA-1) apply directly to
development, coding, and validation of safety and non-safety computer
models used in cleanup decisions. EM has interpreted and applied these
directives and accompanying standards to develop its corporate quality
program, which includes Software Quality Assurance and therefore
computer model development.
We agree with GAO that clarification of specific quality assurance
requirements for computer models Department-wide may be needed. Also,
EM recognizes that there may be instances where oversight of
contractor development and use of computer models can fall short of
quality program expectations. However, EM continuously makes use of
its oversight, corrective action, and lessons learned management
systems to correct these deficiencies. The identification of specific
implementation issues within a quality program does not necessarily
indicate the program is not effective and/or functional. EM shared a
description of its oversight, corrective action, and lessons learned
management systems prior to issuance of GAO's draft report.
Regarding the selection of models used in cleanup decision making, DOE
must adhere to compliance agreements and has a long history of
collaborating with its environmental regulators at the Federal and
state level, other government agencies, tribal nations, and local
government and stakeholders in environmental cleanup decision making.
This includes determining and applying appropriate models and
considering other criteria as mandated by regulation and state-of-the-
art practices in selecting cleanup remedies. These processes have been
compliant with applicable laws and regulations, and the results have
been noteworthy”the Comprehensive Environmental Response,
Compensation, and Liability Act-mandated Five-Year Reviews and DOE
Order 435.1 Radioactive Waste Management-required Annual Summaries
conclude that implemented remedies and performance of low-level
radioactive waste disposal facilities at DOE EM sites are protective
of human health and the environment.
As noted in your report, EM has begun several initiatives to improve
our modeling efforts. Notable among these is the Advanced Simulation
Capability for Environmental Management effort. As discussed with the
GAO, this effort aims at developing improved simulation capabilities
for consistent applications to EM cleanup work across the complex. The
team developing this software consists of the world's foremost nuclear
waste scientists and high performance computing modelers, and includes
participants from eight U.S. National Laboratories. A second
initiative that EM has completed is the purchase and distribution of
150 copies of three (3) software-related national consensus standards
from the American Nuclear Society. DOE hosted a Workshop ("Waste
Processing Models: Material Properties Standards and Software V&V
Training Workshop," November 30-December 1, 2010) where the standards
were distributed and modelers/software developers received training on
their application.
With respect to Information Technology (IT) and software management,
the Principal Deputy Assistant Secretary for EM issued a memorandum to
EM Headquarters and Field leaders on August 13, 2009, subject,
"Information 'Technology Investments in the Office of Environmental
Management." This memorandum established that all requests or
requirements for IT investments must be submitted to the Director,
Office of Corporate Information Technology, for review and approval
prior to initiating, planning, or implementation activities. While
this direction was not specific to computer modeling technologies,
they were clearly included in the scope given that they are IT
investments. Further, this policy document was shared with GAO prior
to the issuance of GAO's draft report, and it is not acknowledged in
the draft.
Additionally, EM Headquarters developed and implemented an IT
Governance process in 2010 to enable EM to more effectively manage IT
with the goals of reducing duplication between EM sites to achieve
maximum cost efficiency, promoting consistency, and sharing lessons
learned across the nuclear weapons complex to enable our cleanup
mission.
We believe that the recommendations made in the draft GAO report will
strengthen our modeling efforts. Provided in Enclosure 1 is our
response to the GAO recommendations, including our proposed path
forward. We are also submitting specific comments in Enclosure 2 that
provide clarification to technical and factual information for your
consideration in preparing your final report.
If you have any questions, please contact Ms. Yvette T. Collazo,
Director, Office of Technology Innovation and Development, at (202)
586-5280.
Sincerely,
Signed by:
Ines R. Triay:
Assistant Secretary for Environmental Management:
2 Enclosures:
[End of letter]
Enclosure 1:
U.S. Department of Energy:
Office of Environmental Management:
Response to GAO Recommendations for Executive Action:
GAO-11-143 - "Nuclear Waste: DOE Needs a Comprehensive Strategy and
Guidance on Computer Models that Support Environmental Cleanup
Decisions" Recommendations for Executive Action (Page 23 of Draft GAO-
11-143):
To help the Office of Environmental Management (EM) increase
confidence in the quality of information provided to the public and
its stakeholders resulting from the use of computer modeling, we
recommend the Secretary of Energy take the following three actions:
* Clarify specific quality assurance requirements for computer models
used to analyze the potential effectiveness of cleanup alternatives,
assess the performance of selected cleanup activities, and assist in
planning and budgeting cleanups.
* Ensure that the models are assessed for compliance with these
requirements.
* Develop a comprehensive strategy and guidance for the management of
computer models to promote consistency, reduce duplication, and ensure
sharing of lessons learned.
Recommendation 1: Clarify specific quality assurance requirements for
computer models used to analyze the potential effectiveness of cleanup
alternatives, assess the performance of selected cleanup activities,
and assist in planning and budgeting cleanups.
Concur. Resolution of this issue will require EM to coordinate with
additional program offices within the Department because EM is not
charged with writing software quality assurance (SQA) policy. The two
technical areas that are of concern regarding model quality assurance
are: (1) Information Technology (IT) project management; and (2) SQA
for both nuclear, and non-nuclear, facilities and applications. The
technical leads for each activity within DOE are described below.
1. Software and IT Systems Project Management: Within EM, the Office
of Corporate Information Technology has the primary responsibility to
ensure that EM IT is acquired and that information resources are
managed in a manner consistent with statutory, regulatory, and
Departmental requirements and priorities for IT systems. The EM Office
of Corporate Information Technology works jointly with the DOE Chief
Information Officer (CIO) in the facilitation of DOE Orders and OMB
requirements for IT, so that information resources can be utilized
effectively and efficiently.
EM is committed to working with the DOE CIO to make sure that IT
project management requirements are clear, communicated to EM field
sites, and implemented in an appropriate manner by EM Headquarters and
field personnel. The DOE CIO has developed an Information Technology
Project Guide (Guide 413.3-14) which defines guidelines for project
implementation. In addition, the EM Office of Corporate Information
Technology has created a separate EM IT Projects Guide that is
specific to EM which incorporates all the requirements from the DOE
CIO IT Project Guide.
2. Software Quality Assurance: The responsible DOE Program Office is
the Office of Health Safety and Security (HSS), Office of Quality
Assurance Policy and Assistance. From the HSS website [hyperlink,
http://www.hss.doe.gov/nuclearsafety/hs23.html], the mission of the
Office of Quality Assurance Policy and Assistance is as follows:
"The Office of Quality Assurance Policy and Assistance establishes and
maintains the quality assurance (QA) policies, requirements and
guidance for the Department and serves as DOE's corporate resource to
ensure that products and services meet or exceed the Department's
quality objectives. The Office provides assistance to Departmental
elements and contractors in the interpretation and implementation of
DOE quality assurance requirements and in the resolution of QA-related
issues."
The EM Office of Standards and Quality Assurance will work with DOE
HSS to make sure software QA policy/guidance is clear. Where
policy/guidance is unclear, or does not exist, EM is committed to
working with DOE HSS to clarify QA requirements and communicate this
guidance to EM Headquarters and field sites.
Recommendation 2: Ensure that the models are assessed for compliance
with these requirements.
Concur. EM is committed to ensuring that models developed in the field
and at EM Headquarters comply with the directives of DOE, and relevant
national consensus standards. The EM Office of Standards and Quality
Assurance and Office of Corporate Information Technology will work
closely with DOE HSS and DOE CIO to assure that EM is compliant with
Departmental directives. In addition, EM will review and, where
needed, develop additional SQA oversight criteria to ensure computer
models that have been or are to be developed within EM comply with
Departmental directives and are implemented appropriately at all DOE
EM facilities.
Recommendation 3: Develop a comprehensive strategy and guidance for
the management of computer models to promote consistency, reduce
duplication, and ensure sharing of lessons learned.
Concur. The EM Office of Corporate Information Technology will follow-
up to ensure that every computer modeling IT investment is documented
through the OMB-required Capital Planning and Investment Control
process across the EM complex. We plan to conduct a survey to ensure
that all the modeling tools in use in EM are included, including those
referenced in this report by July 31, 2011. We will also reissue the
August 13, 2009, memorandum on Information Technology Investments in
EM by February 28, 2011. EM will also process the current slate of
computer models through our IT Governance process with the goal of
streamlining, where appropriate, by December 30, 2011.
[End of section]
Appendix IV: GAO Contact and Staff Acknowledgments:
GAO Contact:
Gene Aloise, (202) 512-3841 or aloisee@gao.gov:
Staff Acknowledgments:
In addition to the individual named above, Ryan T. Coles, Assistant
Director; Ivelisse Aviles; Mark Braza; Dan Feehan; Nancy Kintner-
Meyer; Jonathan Kucskar; Mehrzad Nadji; Kathryn Pedalino; Thomas C.
Perry; and Benjamin Shouse made key contributions to this report.
[End of section]
Related GAO Products:
Nuclear Waste: Actions Needed to Address Persistent Concerns with
Efforts to Close Underground Radioactive Waste Tanks at DOE's Savannah
River Site. [hyperlink, http://www.gao.gov/products/GAO-10-816].
Washington, D.C.: September 14, 2010.
Recovery Act: Most DOE Cleanup Projects Appear to Be Meeting Cost and
Schedule Targets, but Assessing Impact of Spending Remains a
Challenge. [hyperlink, http://www.gao.gov/products/GAO-10-784].
Washington, D.C.: July 29, 2010.
Department of Energy: Actions Needed to Develop High-Quality Cost
Estimates for Construction and Environmental Cleanup Projects.
[hyperlink, http://www.gao.gov/products/GAO-10-199]. Washington, D.C.:
January 14, 2010.
Nuclear Waste: Uncertainties and Questions about Costs and Risks
Persist with DOE's Tank Waste Cleanup Strategy at Hanford. [hyperlink,
http://www.gao.gov/products/GAO-09-913]. Washington, D.C.: September
30, 2009.
Department of Energy: Contract and Project Management Concerns at the
National Nuclear Security Administration and Office of Environmental
Management. [hyperlink, http://www.gao.gov/products/GAO-09-406T].
Washington, D.C.: March 4, 2009.
Nuclear Waste: DOE Lacks Critical Information Needed to Assess Its
Tank Management Strategy at Hanford. [hyperlink,
http://www.gao.gov/products/GAO-08-793]. Washington, D.C.: June 30,
2008.
Hanford Waste Treatment Plant: Department of Energy Needs to
Strengthen Controls over Contractor Payments and Project Assets.
[hyperlink, http://www.gao.gov/products/GAO-07-888]. Washington, D.C.:
July 20, 2007.
Nuclear Waste: DOE Should Reassess Whether the Bulk Vitrification
Demonstration Project at Its Hanford Site Is Still Needed to Treat
Radioactive Waste. [hyperlink,
http://www.gao.gov/products/GAO-07-762]. Washington, D.C.: June 12,
2007.
Hanford Waste Treatment Plant: Contractor and DOE Management Problems
Have Led to Higher Costs, Construction Delays, and Safety Concerns.
[hyperlink, http://www.gao.gov/products/GAO-06-602T]. Washington,
D.C.: April 6, 2006.
Nuclear Waste: Absence of Key Management Reforms on Hanford's Cleanup
Project Adds to Challenges of Achieving Cost and Schedule Goals.
[hyperlink, http://www.gao.gov/products/GAO-04-611]. Washington, D.C.:
June 9, 2004.
Nuclear Waste: Challenges to Achieving Potential Savings in DOE's High-
Level Waste Cleanup Program. [hyperlink,
http://www.gao.gov/products/GAO-03-593]. Washington, D.C.: June 17,
2003.
Nuclear Waste: Department of Energy's Hanford Tank Waste Project--
Schedule, Cost, and Management Issues. [hyperlink,
http://www.gao.gov/products/GAO-RCED-99-13]. Washington, D.C.: October
8, 1998.
[End of section]
Footnotes:
[1] 42 U.S.C. § 9601 et seq.
[2] 42 U.S.C. § 4321 et seq.
[3] EM cleanup activities are also subject to the requirements of the
act commonly known as the Resource Conservation and Recovery Act (42
U.S.C. § 6901 et seq.). Decisions made under this act were not
assessed in this report.
[4] DOE, Radioactive Waste Management, DOE O 435.1 (Washington, D.C.,
July 9, 1999).
[5] To meet the requirements of DOE O 435.1, DOE completes performance
assessments and composite analyses. Performance assessments are
required for specific waste management decisions, while composite
analyses are performed to evaluate the cumulative impacts of waste
management and cleanup actions at a DOE site. Both serve to provide a
reasonable expectation that human health and environmental protection
performance objectives will be met.
[6] Among the cleanup activities Hanford and SRS must address are the
treatment and disposal of millions of gallons of highly radioactive
waste stored in aging and leak-prone underground tanks and removal,
immobilization, or monitoring of radioactive and hazardous
contamination that has migrated through the soil into the groundwater,
posing a threat to human health and the environment. Other activities
include tearing down buildings and removing and disposing of
contaminated soil.
[7] DOE, Draft Tank Closure and Waste Management Environmental Impact
Statement for the Hanford Site, DOE/EIS-0391 (Washington, D.C.,
October 2009). The draft environmental impact statement is scoped to
evaluate the Fast Flux Test Facility, Waste Management, and Tank
Closure, and includes analysis of several alternatives for tank
closure that include, for example, emptying and removing the tanks
from the ground; or emptying the tanks, leaving the tanks in the
ground, and filling them with grout or other material.
[8] DOE, Radioactive Waste Management, DOE O 435.1 (Washington, D.C.,
July 9, 1999).
[9] Savannah River Remediation, LLC, "Performance Assessment for the F-
Tank Farm at the Savannah River Site," prepared for DOE under Contract
No. DE-AC09-09SR22505, SRS-REG-2007-00002 (Aiken, S.C., Mar. 31,
2010). A tank farm is a group of tanks buried side by side in the
ground. In addition to the tanks themselves, tank farms also contain
equipment such as lines and pumps for transferring waste between
tanks, equipment for monitoring heat and chemical reactions inside the
tanks, instruments to measure temperature and tank waste levels, and
other support facilities. Although SRS's F-Tank Farm originally
contained 22 underground liquid radioactive waste tanks, 2 of these
tanks have already been closed.
[10] Savannah River Remediation, LLC, "Liquid Waste System Plan,
Revision 15," prepared for DOE under Contract No. DE-AC09-09SR22505
(Aiken, S.C., Jan. 11, 2010).
[11] DOE EM Headquarters imposes quality assurance through its
Corporate Quality Assurance Program which is, according to DOE, based
on law, DOE directives, national consensus standards, and EM quality
management expectations. The program allows for a graded approach to
quality assurance, specifying additional requirements for software
that relates to nuclear safety.
[12] DOE, Quality Assurance, DOE Order 414.1C (Washington, D.C., June
17, 2005). DOE Order 414.1 was first approved in November 1998.
Although some of the modeling we reviewed was performed as far back as
the early 2000s, DOE Order 414.1 was first approved in 1998 and
applied to that modeling. In addition to DOE Order 414.1C, EM's
quality assurance program is derived from 10 C.F.R. § 830 and EM
quality management expectations. DOE refers to its system of quality
assurance policies and orders as "directives." DOE generally imposes
its quality directives on contractors by inclusion in contracts.
[13] American Society of Mechanical Engineers, "Quality Assurance
Requirements for Nuclear Facility Applications," NQA-1-2000, (New
York, N.Y., May 2001).
[14] The types of assessments that DOE provided ranged from EPA and
state regulator comments on draft environmental impact statements, to
internal quality assessment reviews conducted by contractors, to
general quality assurance reviews that DOE conducted of individual
contractors.
[15] Savannah River Remediation, LLC, "Performance Assessment for the
F-Tank Farm at the Savannah River Site," prepared for DOE under
Contract No. DE-AC09-09SR22505, SRS-REG-2007-00002 (Aiken, S.C., Mar.
31, 2010).
[16] DOE, External Technical Review for Evaluation of System Level
Modeling and Simulation Tools in Support of SRS Liquid Waste Process
(June 2009) and DOE, External Technical Review for Evaluation of
System Level Modeling and Simulation Tools in Support of Hanford Site
Liquid Waste Process (September 2009).
[17] KD Auclair & Associates, LLC, Independent Review of the Draft
Tank Closure and Waste Management Environmental Impact Statement
(Benton City, Wash., March 2010).
[18] GAO, Ways to Improve Management of Federally Funded Computerized
Models, [hyperlink, http://www.gao.gov/products/LCD-75-111]
(Washington, D.C.: Aug. 25, 1976).
[19] National Research Council, Models in Environmental Regulatory
Decision Making, (Washington, D.C., 2007).
[20] EPA, Guidance on the Development, Evaluation, and Application of
Environmental Models, EPA/100/K-09/003 (Washington, D.C., March 2009).
[21] DOE's Office of Civilian Radioactive Waste Management was
terminated on September 30, 2010.
[End of section]
GAO's Mission:
The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting
its constitutional responsibilities and to help improve the performance
and accountability of the federal government for the American people.
GAO examines the use of public funds; evaluates federal programs and
policies; and provides analyses, recommendations, and other assistance
to help Congress make informed oversight, policy, and funding
decisions. GAO's commitment to good government is reflected in its core
values of accountability, integrity, and reliability.
Obtaining Copies of GAO Reports and Testimony:
The fastest and easiest way to obtain copies of GAO documents at no
cost is through GAO's Web site [hyperlink, http://www.gao.gov]. Each
weekday, GAO posts newly released reports, testimony, and
correspondence on its Web site. To have GAO e-mail you a list of newly
posted products every afternoon, go to [hyperlink, http://www.gao.gov]
and select "E-mail Updates."
Order by Phone:
The price of each GAO publication reflects GAO‘s actual cost of
production and distribution and depends on the number of pages in the
publication and whether the publication is printed in color or black and
white. Pricing and ordering information is posted on GAO‘s Web site,
[hyperlink, http://www.gao.gov/ordering.htm].
Place orders by calling (202) 512-6000, toll free (866) 801-7077, or
TDD (202) 512-2537.
Orders may be paid for using American Express, Discover Card,
MasterCard, Visa, check, or money order. Call for additional
information.
To Report Fraud, Waste, and Abuse in Federal Programs:
Contact:
Web site: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]:
E-mail: fraudnet@gao.gov:
Automated answering system: (800) 424-5454 or (202) 512-7470:
Congressional Relations:
Ralph Dawn, Managing Director, dawnr@gao.gov:
(202) 512-4400:
U.S. Government Accountability Office:
441 G Street NW, Room 7125:
Washington, D.C. 20548:
Public Affairs:
Chuck Young, Managing Director, youngc1@gao.gov:
(202) 512-4800:
U.S. Government Accountability Office:
441 G Street NW, Room 7149:
Washington, D.C. 20548:
