Nuclear Regulatory Commission: Oversight of Nuclear Power Plant  
Safety Has Improved, but Refinements Are Needed (27-SEP-06,	 
GAO-06-1029).							 
                                                                 
The Nuclear Regulatory Commission (NRC) is responsible for	 
overseeing the nation's 103 commercial nuclear power plants to	 
ensure they are operated safely. The safety of these plants has  
always been important, since an accident could release harmful	 
radioactive material. NRC's oversight has become even more	 
critical as the potential resurgence of nuclear power is	 
considered. NRC implemented a new Reactor Oversight Process (ROP)
in 2000 to address weaknesses in its oversight of nuclear plant  
safety. In this report, GAO reviewed (1) how NRC oversees nuclear
power plants, (2) the results of the ROP over the past several	 
years, and (3) the status of NRC's efforts to improve the ROP. To
complete this work, GAO analyzed programwide information,	 
inspection results covering 5 years of ROP operations, and	 
detailed findings from a nonprobability sample of 11 plants.	 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-06-1029					        
    ACCNO:   A61465						        
  TITLE:     Nuclear Regulatory Commission: Oversight of Nuclear Power
Plant Safety Has Improved, but Refinements Are Needed		 
     DATE:   09/27/2006 
  SUBJECT:   Evaluation criteria				 
	     Federal regulations				 
	     Inspection 					 
	     Noncompliance					 
	     Nuclear powerplant safety				 
	     Nuclear powerplants				 
	     Performance measures				 
	     Radiation safety					 
	     Risk assessment					 
	     Safety regulation					 
	     Safety standards					 
	     Policies and procedures				 
	     NRC Reactor Oversight Process			 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-06-1029

     

     * Report to Congressional Requesters
          * September 2006
     * NUCLEAR REGULATORY COMMISSION
          * Oversight of Nuclear Power Plant Safety Has Improved, but
            Refinements Are Needed
     * Contents
          * Results in Brief
          * Background
          * NRC Uses Various Tools and Takes a Risk- Informed and Graded
            Approach to Ensuring the Safety of Nuclear Power Plants
               * NRC Collects Information about Plant Performance from
                 Physical Inspections and Quantitative Measures Reported by
                 the Licensees
               * NRC Uses Its Action Matrix to Categorize Plant Performance
                 and Apply Increased Oversight in a Graded Fashion
          * NRC Has Identified Low Risk Problems at Nuclear Power Plants,
            Resulting in Increased Oversight for Varying Periods
          * NRC Is Addressing Weaknesses in Various Areas of Its Oversight
            Process, but More Effort Is Needed
               * NRC Is Taking Action to Improve Various Areas of Its
                 Oversight Process
               * NRC Is Taking Its First Major Step to Address a Significant
                 ROP Weakness in the Area of Safety Culture
          * Conclusions
          * Recommendations for Executive Action
          * Agency Comments
     * Key Safety-Related Events at the Salem and Hope Creek Nuclear Power
       Plants from 2000 to 2006
          * Summary of Key Safety-Related Events at Salem and Hope Creek,
            April 2000 to June 2006
               * White Performance Indicator for Unplanned Scrams
               * White Performance Indicator for Unplanned Changes in Reactor
                 Power
               * Substantive Cross-Cutting Issue in Problem Identification
                 and Resolution
               * White Inspection Finding for Failure of a Part on an
                 Emergency Diesel Generator
               * White Inspection Finding for the Failure of a Rotating
                 Screen That Is Part of the Station Service Water System at
                 Hope Creek
               * White Inspection Finding for the Failure of a Drain Line in
                 the Moisture Separator System at Hope Creek
               * Increased Oversight to Address Safety-Conscious Work
                 Environment Problems
     * Scope and Methodology
     * Nuclear Power Plant Performance Data on the Basis of the Results of
       NRC's Reactor Oversight Process, 2001 Through 2005
     * Comments from the Nuclear Regulatory Commission
     * GAO Contact and Staff Acknowledgments
     * PDF6-Ordering Information.pdf
          * Order by Mail or Phone

Report to Congressional Requesters

September 2006

NUCLEAR REGULATORY COMMISSION

Oversight of Nuclear Power Plant Safety Has Improved, but Refinements Are
Needed

Contents

Tables

Figures

September 27, 2006Letter

Congressional Requesters

The Nuclear Regulatory Commission (NRC) is responsible for overseeing the
safe operation of the nation's 103 operating commercial nuclear power
plants, which provide about 20 percent of U.S. electricity. The safety of
these plants, which are located at 65 sites in 31 states, has always been
important, since an accident could result in the release of radioactive
material and potentially harm public health and the environment. NRC's
oversight has become even more critical as the Congress and the nation
consider the potential resurgence of nuclear power in helping to meet the
nation's growing energy needs. No new orders for plants have been placed
since the 1979 accident at the Three Mile Island plant in Pennsylvania.
However, in the face of concerns about energy security, global warming,
aging plants, and the ever increasing need for energy to fuel the nation's
economy, nuclear power is resurfacing as a principal option. An accident,
even on a relatively small scale, could threaten public confidence in
nuclear power just as it begins to emerge from the shadows of the Three
Mile Island accident. It is critical that NRC be able to ensure that
nuclear power plants are operated safely and that public confidence about
their safety is high.

NRC is responsible for issuing regulations, licensing and overseeing
plants, and requiring necessary actions to protect public health and
safety, up to and including shutting down a plant, if it is not meeting
licensing conditions and it poses an undue risk to public health and
safety. Plant operators are responsible for safely operating their plants
in accordance with their licenses. Prior to 2000, NRC's oversight of
plants' compliance, which is a critical part of the agency's regulatory
program, was criticized because it did not always focus on the most
important safety issues, and because some activities were redundant,
inefficient, and overly subjective. While its new process-which NRC refers
to as the Reactor Oversight Process (ROP)-is similar to its prior process
in that the oversight activities largely consist of physical plant
inspections, the inspections now focus on more important safety issues.
NRC's goal is to ensure that its oversight of plants' safety performance
is objective, predictable, understandable, and conducted openly to inform
the public and maintain confidence about plant safety. The unexpected
discovery in March 2002 of extensive corrosion and a pineapple-size cavity
in the reactor vessel head-one of the vital barriers preventing a
radioactive release-at the Davis-Besse nuclear power plant in Ohio led NRC
to reexamine its safety oversight and other regulatory processes to
determine how such corrosion could have been missed. On the basis of the
lessons learned from that event, NRC made several changes to the ROP. NRC
also has assessed the ROP annually since it was implemented in 2000 by
obtaining feedback from the industry and other stakeholders, such as
public interest groups.

In this report, we examined (1) how NRC oversees nuclear power plants to
ensure that they are operated safely, (2) the results of the ROP over the
past several years, and (3) the status of NRC's efforts to improve the
ROP.1 In addition, this report provides details on recent safety-related
events at the Salem and Hope Creek nuclear power plants in New Jersey and
on NRC's and the licensee's actions in response to these events (see app.
I).

To examine how NRC oversees plants, we reviewed the various tools and
processes that comprise the ROP. In this regard, we reviewed NRC's
policies, inspection manuals, and other guidance documents; interviewed
NRC headquarters and regional program officials and regional and on-site
inspectors; visited the Salem and Hope Creek nuclear power plants; and
attended several public meetings covering various nuclear power plant
oversight topics. To examine the results of the ROP over the past several
years, we reviewed the number and types of inspection findings NRC issued,
along with other performance measures or indicators it collected from the
plants, and the level of oversight it provided as a result of its
findings. We analyzed NRC data on nuclear plant safety for 2001 through
2005, the years since implementation of the ROP for which data were
available for the full year, and discussed our analysis with NRC program
officials. We assessed the reliability of these data and determined that
they were sufficiently reliable for the purposes of our report. To examine
the status of NRC's efforts to improve the ROP, we analyzed NRC documents,
including annual self-assessment reports; interviewed officials from NRC,
including a former Commission Chairman who is largely credited with
leading the development of the ROP, and outside stakeholder groups,
including the Nuclear Energy Institute (NEI), the Union of Concerned
Scientists (UCS), and Greenpeace; and attended several key public meetings
covering proposed changes to oversight procedures. We also reviewed
various external evaluations of the ROP, including our prior reports and
those of the Advisory Committee on Reactor Safeguards (ACRS)2 and the NRC
Office of the Inspector General. In addition, we selected a nonprobability
sample of 6 nuclear power sites (totaling 11 plants, including Salem and
Hope Creek) that spanned each of NRC's four regions and represented
varying levels of plant performance and NRC oversight since 2000. We
reviewed relevant inspection reports and assessment documents and
interviewed NRC and industry officials at each site to examine how NRC
applies the ROP to identify and correct safety problems. Appendix II
presents a detailed discussion of our scope and methodology. We conducted
our work from July 2005 through July 2006 in accordance with generally
accepted government auditing standards.

Results in Brief

NRC uses various tools and takes a risk-informed and graded approach to
ensuring the safety of nuclear power plants. These tools include physical
inspections of plants' equipment and records and quantitative measures or
indicators of plant performance, such as the number of unplanned reactor
shutdowns and the reliability of alert and notification system sirens that
notify residents living near the plant in the event of an accident. NRC
uses a risk-informed approach-that is, an approach that considers safety
significance in selecting the equipment or operating procedures to be
inspected-to apply these tools. NRC inspectors conduct baseline
inspections of plant operations almost continuously at each nuclear power
plant site. When NRC becomes aware of a performance problem at a plant, it
uses a process to assign the inspection finding one of four colors to
reflect the finding's risk significance, which is set on the basis of
measures that reflect the potential health effects that could occur from
radiological exposure. Green inspection findings equate to very low risk
significance, while white, yellow, and red colors represent increasing
levels of risk, respectively. In response to greater-than-green (white,
yellow, or red) inspection findings, NRC conducts supplemental
inspections, which expand the scope of baseline inspections and review (1)
the extent of the problem, (2) the sufficiency of the licensee's
evaluation of the root cause of the problem, and (3) the licensee's
proposed corrective actions in response to the identified performance
problem. NRC conducts special inspections to investigate specific safety
incidents or events-such as reactor shutdowns due to equipment
failures-that are of particular interest to NRC because of their potential
significance to safety. Nuclear power plants also self-report on their
safety performance, using performance measures or indicators in quarterly
reports that they submit to NRC. The plants' reports are verified by NRC's
on-site inspectors. On the basis of the number and risk significance of
inspection findings and performance indicators, NRC places each plant into
one of five oversight categories on its action matrix, which correspond to
graded or increasing levels of oversight, largely consisting of
supplemental inspections that increase in the breadth and depth of issues
covered as plants move into higher oversight categories. NRC provides an
overall assessment of each plant's performance through assessment letters
issued to plants at the end of each 6-month period describing their
specific performance and the level of oversight that will result. In
addition, NRC has other mechanisms to make available its oversight
results, such as an Internet Web site devoted to the ROP that provides
detailed summaries of each plant's performance.

Since 2001, the ROP has resulted in more than 4,000 inspection findings
concerning nuclear power plant licensees' failure to fully comply with NRC
regulations and industry standards for safe plant operation, and NRC has
subjected more than 75 percent (79) of the 103 operating plants to
increased oversight for varying periods. (See app. III for additional
site-specific plant data.) About 97 percent of the inspection findings
were green, meaning they were for actions or failures NRC considered
important to correct but of very low significance to overall safe plant
operations. For example, a finding of very low risk significance was
issued at one plant after a worker failed to wear the proper radiation
detector, and at another plant because the operator failed to properly
evaluate and approve the storage of flammable materials in the vicinity of
safety-related equipment. The other 3 percent (98) of the inspection
findings were greater-than-green. Most of these findings (86 of the 98)
were white, meaning they were considered to be of low-to-moderate risk
significance. The other 12 inspection findings, or less than 1 percent,
were of the highest levels of significance to safety (yellow and red). For
example, NRC issued a finding of the highest risk significance (red) at
one plant after a steam generator tube failed, causing an increased risk
of the release of radioactive material. In the area of performance
indicators, there were 156 instances out of more than 30,000 reports, or
less than 1 percent, in which data reported for individual indicators were
outside of NRC's acceptable performance category. On the basis of
greater-than-green inspection findings and performance indicators, NRC has
subjected more than 75 percent (79) of the 103 operating plants to
oversight beyond the baseline inspections for varying periods. Most of
these plants received the lowest level of increased oversight, consisting
of a supplemental inspection to follow up on the corrective actions taken
for performance problems identified through the issuance of 1 or 2 white
inspection findings or performance indicators. Over the past 5 years, 5
plants have been subjected to the highest level of NRC oversight that
still allows continued operations. Plants in this category were generally
subjected to this higher oversight for long periods due to the more
intensive supplemental inspections conducted by NRC and the more systemic
nature of the plants' performance problems and subsequent corrective
actions NRC expected the licensees to take. NRC inspectors at the plants
we reviewed indicated that when plant performance declines, it is often
the result of ineffective corrective action programs, problems related to
human performance, or complacent management. In assessing the results of
the ROP between 2001 and 2005, we found an association between poorer
performing plants and deficiencies in the plants' human performance and
problem identification and resolution programs.

NRC has improved its oversight process in various areas, but continued
efforts will be needed to address other shortcomings or opportunities for
improvement, particularly in improving the agency's ability to identify
and address early indications of declining plant safety performance. NRC
has made several improvements, largely in response to independent reviews
and feedback from stakeholders, including its regional and on-site
inspectors, usually obtained during NRC's annual self-assessment of its
oversight process. These improvements include better focusing its
inspections on those areas most important to safety, reducing the time
needed to determine the risk significance of inspection findings, and
modifying the way that some performance indicators are measured to improve
their quality. NRC also is assessing whether it needs to modify its
oversight, including developing additional inspection procedures, as a
result of some problems that have surfaced in areas not fully inspected by
NRC, such as the recent discovery of groundwater contamination from
radioactive materials at a number of sites. For the most part, NRC
considers these efforts to be refinements to its oversight process, rather
than significant changes. One important shortcoming in the ROP that we and
others have found is that it is not as effective as it could be in
identifying and addressing early indications of deteriorating safety at
nuclear power plants before problems develop. In response to this concern,
NRC recently undertook a major initiative to improve its ability to
address plants' safety culture-that is, the organizational characteristics
that ensure issues affecting nuclear plant safety receive the attention
their significance warrants. NRC and others have long recognized that
safety culture attributes, such as attention to detail, adherence to
procedures, and effective corrective and preventative actions, have a
significant impact on a plant's safety performance, and that the lessening
of these attributes can indicate a decline in safety performance before
safety problems occur. However, NRC has been somewhat slow to react and
only recently modified its oversight process by redefining and increasing
its focus on cross-cutting safety issues-issues that comprise many of the
elements of safety culture-and developing new requirements under the ROP
to more directly assess safety culture at poorer performing plants.
However, some of NRC's changes have been controversial. While some
industry officials have expressed concern that the changes could introduce
undue subjectivity to NRC's oversight, given the difficulty in measuring
these often intangible and complex concepts, other stakeholders believe
the changes will provide NRC with better tools to identify safety culture
issues at plants and thus provide earlier indications of declining safety
performance. NRC has been reluctant to incorporate safety culture into the
ROP because it considered this type of activity a management function, and
NRC did not believe that it should be directly involved in managing
licensees' plants. NRC program officials view these changes as the
beginning step in an incremental approach and acknowledge that they will
need to assess the changes they made to the ROP to determine their
effectiveness in better allowing inspectors to detect deteriorating safety
conditions at plants before significant safety events occur.

Given the importance of this initiative to the ROP's effectiveness, we are
recommending that NRC aggressively monitor; evaluate; and, if needed,
implement additional measures to increase the effectiveness of its safety
culture changes. In line with NRC's desire to make the ROP an open
process, we are also recommending that the agency make available
additional information on plants' safety culture to the public and its
other stakeholders to provide a more comprehensive picture of plant
performance. In commenting on a draft of this report, NRC generally agreed
with our findings, conclusions, and recommendations. Appendix IV contains
a reproduction of NRC's letter.

Background

NRC is an independent agency of over 3,200 employees established by the
Energy Reorganization Act of 1974 to regulate civilian-that is,
commercial, industrial, academic, and medical-use of nuclear materials.
NRC is headed by a five-member Commission. The President appoints the
Commission members, who are confirmed by the Senate, and designates one of
them to serve as the Chair and official spokesperson. The Commission as a
whole formulates policies and regulations governing nuclear reactor and
materials safety, issues orders to licensees, and adjudicates legal
matters brought before it. NRC's staff work out of its headquarters office
in Rockville, Maryland; out of its four regional offices; and at each of
the operating commercial nuclear power plant sites (see fig. 1). NRC's
Office of Nuclear Reactor Regulation provides overall direction for the
oversight process and the Office of Enforcement is responsible for
ensuring that appropriate enforcement actions are taken when performance
issues are identified. NRC's regional offices are responsible for
implementing the ROP, along with the inspectors who work directly at each
of the nuclear power plant sites.

Figure 1: NRC Regions and Operating Nuclear Power Plant Sites in the
United States

Nuclear power plant licensees have the primary responsibility for safely
operating their plants in accordance with their licenses and NRC
regulations. The plants have many physical structures, systems, and
components, and licensees have numerous activities under way 24 hours a
day to ensure that plants operate safely. NRC relies on, among other
things, its on-site inspectors to assess plant conditions and the
licensees' quality assurance programs, such as those required for
maintenance and problem identification and resolution. Given the numerous
activities going on during complex plant operations, NRC can inspect only
a relatively small sample of the plants' activities. According to NRC, its
focus on the more safety-significant activities is made possible because
safety performance at plants has improved as a result of more than 25
years of operating experience, and because improvements have been made in
the risk assessment tools available to NRC inspectors.

Commercial nuclear power plants are designed according to a
"defense-in-depth" philosophy revolving around redundant, diverse, and
reliable safety systems. For example, two or more key safety components
are put in place so that if one fails, there is another to back it up.
Plants have numerous built-in sensors to monitor important indicators,
such as water temperature and pressure. Plants also have physical barriers
to contain radiation and provide emergency protection. For example,
nuclear fuel is contained in a ceramic pellet to lock in the radioactive
byproducts, the fuel pellets are sealed inside rods made of special
material designed to contain fission products, and the fuel rods are
placed in reactors housed in containment buildings made of several feet of
concrete and steel.

In addition, the nuclear power industry formed an organization, the
Institute of Nuclear Power Operations (INPO), whose mission is to "promote
the highest levels of safety and reliability, to promote excellence, in
the operation of nuclear electric generating plants." INPO provides a
system of personnel training and qualification for all key positions at
nuclear power plants, and workers undergo both periodic training and
assessment. INPO also conducts periodic evaluations of operating plants,
focusing on plant safety and reliability, in the areas of operations,
maintenance, engineering, radiological protection, chemistry, and
training. Licensees make the results of these evaluations available to NRC
for review, and NRC staff use the evaluations as a means to determine
whether its oversight process has missed any performance issues.

Prior to the ROP, NRC conducted Systematic Assessment of Licensee
Performance (SALP) evaluations. SALP evaluations were largely based on
physical plant inspections conducted at each operating plant. Every 12 to
24 months, NRC provided an overall assessment of plant safety performance.
As part of the assessment process, NRC's senior management met to evaluate
plants' performance and develop a "watch list" of those plants requiring
increased regulatory attention. The SALP process was heavily criticized by
the industry and other internal and external stakeholders, however, for
being inconsistently applied among NRC regions and for lacking clear and
consistent responses once issues were identified. In 1997 and 1998, we
reported that NRC's oversight needed substantial revisions to achieve its
purpose as an early-warning tool, and that NRC did not consistently apply
the SALP across the industry.3 We found the inconsistency could be
attributed, in part, to a lack of specific criteria, the subjective nature
of the process, and an ineffective process for ensuring that the licensees
maintain competent management at their plants.

To address these concerns, NRC undertook a major effort to revise its
oversight process. NRC held a series of public meetings and workshops and
formed several task groups to involve its internal and external
stakeholders, including NEI, UCS, state agencies, and others. In 1999, NRC
conducted a 6-month pilot program to implement the ROP at various sites
across the country. On the basis of the input it received from its
stakeholders and the results from the pilot program, NRC finalized its new
process and implemented the ROP at all plants in April 2000, which was a
significant departure from its prior SALP process.

Soon after the ROP was implemented, NRC faced a number of unanticipated
events- including the attacks of September 11, 2001, and the discovery of
the Davis-Besse reactor vessel head degradation in 2002-that challenged
its ability to complete its baseline inspection activities at all plants
in 2002 and 2003. Therefore, NRC staff implemented "coping strategies,"
which consisted of increasing the use of overtime and scheduling
flexibility for its inspectors, reducing the level of effort for some of
its inspection procedures, reducing some of the inspection preparation
time, and deferring some inspections, among other things. A 2004 audit by
the NRC Office of the Inspector General found that the resource challenges
were largely due to changes in NRC's staffing policy, a hiring policy
change, an increase in inspection activities due to unanticipated events
such as Davis-Besse, and a loss of qualified inspectors.4 NRC increased
its inspection resources by 9 percent in 2004, and then by another 5
percent in 2005, and reported that it was able to fully implement its
baseline inspection program at all plants for both years. NRC reports show
that resources expended in 2005 were almost 20 percent higher than those
expended in 2002, the lowest level of inspection resources devoted to the
ROP since its inception in 2000. According to NRC, the additional
resources expended in 2005 were due, in part, to increased oversight that
was based on lessons learned from the Davis-Besse incident and on the
increased focus on security and emergency preparedness. With its current
resource levels, NRC program officials believe they will be able to
continue to implement all program requirements without the need to employ
coping strategies.

NRC Uses Various Tools and Takes a Risk-Informed and Graded Approach to
Ensuring the Safety of Nuclear Power Plants

NRC's tools to oversee the safe operation of nuclear power plants
generally consist of physical inspections of the various complex plant
equipment and operations, reviews of plant records, and quantitative
measures or indicators of plant performance. These tools are risk-informed
in that they focus on the aspects of operations considered most important
to plant safety. NRC bases its oversight process on the principle and
requirement that licensees have programs in place to routinely identify
and address performance issues without NRC's direct involvement. Thus, an
important aspect of NRC's inspection process is ensuring the effectiveness
of licensee programs designed to identify and correct problems. On the
basis of the number and risk significance of inspection findings and
performance indicators, NRC places each plant into one of five performance
categories on its action matrix, which corresponds to graded, or
increasing, levels of oversight. NRC assesses overall plant performance
and communicates the results to licensees and the public on a semiannual
basis.

NRC Collects Information about Plant Performance from Physical Inspections
and Quantitative Measures Reported by the Licensees

Physical plant inspections are the main tool NRC uses to oversee plant
safety performance. NRC defined specific inspection areas by developing a
list of those elements most critical to meeting the overall agency mission
of ensuring nuclear power plant safety. These safety elements-or key plant
inspection areas-are known as cornerstones. Table 1 summarizes the
objectives of each cornerstone.

Table 1: Key ROP Plant Inspection Areas, or Cornerstones

                                        

         Cornerstone                           Objective                      
Initiating events       Limit the frequency of those events that upset     
                           plant operating stability and challenge critical   
                           safety functions.                                  
Mitigating systems      Ensure the availability, reliability, and          
                           capability of systems that mitigate initiating     
                           events to prevent reactor accidents.               
Barrier integrity       Ensure that physical barriers, such as fuel        
                           cladding and containment structures, protect the   
                           public from radioactive releases caused by         
                           accidents.                                         
Emergency preparedness  Ensure that actions taken by the emergency plan    
                           would provide protection of the public health and  
                           safety during a radiological emergency.            
Occupational radiation  Ensure adequate protection of worker health and    
safety                  safety from exposure to radioactive material       
                           during routine civilian nuclear reactor operation. 
Public radiation safety Ensure adequate protection of public health and    
                           safety from exposure to radioactive material       
                           released into the public domain as a result of     
                           routine civilian nuclear reactor operations.       
Physical protectiona    Provide assurance that the physical protection     
                           system can protect against radiological sabotage.  

Source: NRC.

aThe physical protection cornerstone consists of physical security issues
and, therefore, is outside the scope of this review.

During fiscal year 2005, NRC reported that inspectors spent 411,490 hours
(an average of 77 hours per week at each plant) on plant inspections,
which consist of baseline, supplemental, and special inspections. About 73
percent of this time was devoted to baseline inspections, which are the
minimum level of inspections that all plants receive regardless of
performance, and are conducted on an almost continuous basis. Baseline
inspections are conducted by the two or three NRC inspectors located at
each site and specialists who travel to each site from NRC's regional
offices. These inspections are designed to detect declining safety
performance in each of the cornerstones, and to review licensee
effectiveness at identifying and resolving its safety problems. There are
more than 30 baseline inspection procedures conducted at intervals that
range from quarterly to triennially. These procedures involve both
physical observations of plant activities and reviews of plant reports and
data. Each of the baseline procedures specify a range of sample activities
to inspect. Inspectors then select the type and number of activities to
review on the basis of factors such as the sample activities available;
their risk significance; the amount of time since a particular system or
component was last inspected; and the inspector's judgment, which is based
on information such as reviews of the licensee's corrective action
program, allegations, or plant employee interviews. Risk is factored into
the baseline inspection procedures in the following four ways: (1) areas
of inspection are included in the set of baseline procedures, in part, on
the basis of their risk importance; (2) risk information is used to help
determine the frequency and scope of inspections; (3) the selection of
activities to inspect within each procedure is informed with
plant-specific risk information; and (4) the inspectors are trained in the
use of risk information in planning their inspections. In addition to the
more than 30 baseline inspection procedures, inspectors spend an average
of 750 to 1,100 hours per year (14 to 21 hours per week), depending on the
size of the site, conducting plant status reviews. These reviews are to
ensure that inspectors are aware of plant conditions on a routine basis
and include such activities as reviewing control room activities and
status, attending licensee meetings, and conducting walk-downs of various
plant areas.

When NRC inspectors identify a finding they consider to be more than
minor,5  they use a significance determination process (SDP) to assign the
finding one of four colors to reflect the finding's risk significance. The
SDP assesses how an identified inspection finding increases the risk that
a nuclear accident could occur, or how the finding affects the ability of
the plant safety systems or personnel to prevent such an accident. Risk
thresholds for each color were set on the basis of measures that reflect
the potential health effects that could occur from radiological exposure.6
Green inspection findings equate to very low risk significance, while
white, yellow, and red colors represent increasing levels of risk,
respectively. For greater-than-green (white, yellow, or red) inspection
findings, NRC issues a preliminary color determination after an initial
analysis. It then analyzes any readily available information from the
licensee pertinent to the finding to ensure that the final determination
of risk significance is made with the best available information.

When NRC issues one or more greater-than-green inspection findings at a
plant, it conducts supplemental inspections. Supplemental inspections,
performed by regional staff, expand the scope beyond baseline inspection
procedures and focus on diagnosing the cause of the performance
deficiency. There are three levels of supplemental inspections that are
increasingly expansive in the breadth and depth of their analysis. The
level of supplemental inspection to be carried out depends on the number
and type of performance problem identified. The lowest level of
supplemental inspection assesses the licensee's corrective actions to
ensure they were sufficient in both correcting the problem and identifying
and addressing the root and contributing causes to prevent recurrence. The
second level of supplemental inspection has an increased scope that
includes independently assessing the extent of the condition for both the
specific and any broader performance problems. The highest level of
supplemental inspection is even more comprehensive and includes
determining whether the plant can continue to operate and whether
additional regulatory actions are necessary. The highest level of
supplemental inspection is usually conducted by a multidisciplinary team
of NRC inspectors and may take place over several months. Also, as a part
of this supplemental inspection, NRC inspectors assess the adequacy of the
licensee's overall programs for identifying, evaluating, and correcting
its performance issues, among other things.

NRC conducts special inspections at plants when specific events occur that
are of particular interest to NRC because of their potential safety
significance at the plant or because of potential generic safety concerns
important to all plants. Special inspections determine the cause of the
event and assess the licensee's response to the event. For special
inspections, a team of experts is often formed and an inspection charter
issued that describes the scope of the inspection efforts. At one plant we
reviewed, for example, a special inspection was conducted to investigate
the circumstances surrounding the discovery of leakage from a spent fuel
storage pool.7 Among the objectives of this inspection were to assess the
adequacy of the plant licensee's determination of the source and cause of
the leak, the risk significance of the leakage, and the proposed
strategies to mitigate leakage that had already occurred and repair the
problem to prevent additional leakage.

As part of its inspection process, NRC evaluates all of its findings to
determine if certain elements of plant performance, referred to as
cross-cutting aspects, were a contributing cause to the performance
problem. Cross-cutting aspects represent licensee performance elements
that extend across all of the cornerstones of safety. There are three
cross-cutting aspect areas: (1) problem identification and resolution, (2)
human performance, and (3) a safety-conscious work environment. For
example, in analyzing the failure of a valve to operate properly at one
plant, NRC inspectors determined that plant employees had not followed the
correct maintenance procedures. Thus, NRC concluded that the finding was
associated with the human performance cross-cutting area. Every 6 months,
NRC analyzes all findings issued at each plant during a 12-month
assessment period. If more than three findings have similar causes within
the same cross-cutting area and if NRC is concerned about the licensee's
progress in addressing these issues, it determines that the licensee has a
"substantive" cross-cutting issue. NRC notifies the licensee that it has
opened a substantive cross-cutting issue, and it may ask the licensee to
respond with the corrective actions it plans to take. Also, NRC inspectors
said they provide additional focus on any substantive cross-cutting issues
open through their baseline inspection efforts. For example, one regional
official said that if a substantive cross-cutting issue in the problem
identification and resolution area is identified, it is the region's
practice to increase the frequency of certain baseline inspections that
focus on the licensee's problem identification and resolution processes.

In addition to its various inspections, NRC also collects plant
performance information through its performance indicator program, which
it maintains in cooperation with the nuclear power industry. On a
quarterly basis, each plant submits data for 15 separate performance
indicators8-quantitative measures of plant performance related to safety
in the different aspects of plant operations. For example, one indicator
measures the number of unplanned reactor shutdowns during the previous
four quarters, while another measures the capability of alert and
notification systems that notify residents living near the plant in the
event of an accident. Working with the nuclear power industry, NRC set
thresholds for acceptable performance and assigned colors to each of the
indicators to reflect increasing risk. In contrast to inspection findings,
a green indicator does not indicate a performance deficiency but instead
reflects performance within the acceptable range, while white, yellow, and
red represent decreasing levels of plant performance. NRC inspectors
review and verify the data submitted for each performance indicator
annually through their baseline inspections. If questions arise during the
review and verification process about how to calculate a particular
indicator or what the correct value should be, there is a formal feedback
process to resolve the issue. Similar to its process for conducting
supplemental inspections when greater-than-green inspection findings are
identified, NRC conducts supplemental inspections in response to any
greater-than-green performance indicators.

NRC Uses Its Action Matrix to Categorize Plant Performance and Apply
Increased Oversight in a Graded Fashion

Under the ROP, NRC places each nuclear power plant into one of five
performance categories on its action matrix, which corresponds to graded,
or increasing, levels of oversight.9 The action matrix is NRC's formal
method of determining how much additional oversight-mostly in the form of
supplemental inspections-is required on the basis of the number and risk
significance of inspection findings and performance indicators. (See fig.
2 for an overview of the process that leads to plant placement on the
action matrix.)

Figure 2: NRC's Oversight Process in Determining Plant Placement on the
Action Matrix

The definitions for the categories of the action matrix indicate more
pervasive and systematic declines in licensee performance as a licensee
moves from left to right on the action matrix (see table 2). Also, as the
licensee moves to the right on the action matrix, NRC's response and the
corrective actions it expects of the licensee become more extensive. In
determining a plant's placement on the action matrix, greater-than-green
inspection findings are considered for additional oversight for a minimum
of four quarters, regardless of whether the licensee corrected the
problem, to allow sufficient time to identify additional findings that may
indicate more pervasive performance problems. If a licensee fails to
correct the performance problem within the initial four quarters, the
finding may be held open and considered for additional oversight. NRC
regional officials, with the approval of the Executive Director for
Operations, can also increase or decrease oversight beyond the actions
specified by the action matrix through deviations to the ROP. Deviations
are for rare instances when the regulatory actions dictated by the action
matrix are not appropriate and a more tailored approach is warranted.

Table 2: NRC Reactor Oversight Process Action Matrix

                                        

    Oversight    Lowest    Second level Third level    Highest    Unacceptable 
    category      level                              level (while performance  
                                                       allowing      (plant    
                                                      continued    shutdown)   
                                                     operations)  
Plant       All green   One white    Two or more  Two white    Overall      
performance findings    finding or   white        findings or  unacceptable 
               and         performance  findings or  performance  performance  
               performance indicator,   performance  indicators   due to       
               indicators  or two white indicators   or one       unacceptable 
                           findings or  in one       yellow       margin of    
                           performance  cornerstone, finding or   safety       
                           indicators   or one       performance  
                           in different yellow       indicator in 
                           cornerstones finding or   one          
                                        performance  cornerstone  
                                        indicator,   for five or  
                                        or any three more         
                                        white        quarters, or 
                                        findings or  multiple     
                                        performance  yellow or    
                                        indicators   one red      
                                                     finding or   
                                                     performance  
                                                     indicator    
NRC         Baseline    Baseline     Baseline     Baseline     Order to     
oversight   inspections inspections  inspections  inspections  modify,      
actions     only        and first    and second   and third    suspend, or  
                           level, or    level, or    level, or    revoke       
                           least        more         most         licensed     
                           intensive,   intensive,   intensive,   activities   
                           supplemental supplemental supplemental 
                           inspection   inspection   inspection.a 

Source: NRC.

Note: In addition to the actions listed in this table, increasingly higher
levels of NRC management will meet with the licensee as it moves to the
right on the action matrix.

aFor plants at this oversight level, at a minimum, the licensee and NRC
are to document agreement on the corrective actions the licensee will take
through a performance improvement plan. NRC may also take actions
including making a demand to the licensee for information or issuing an
order up to and including a plant shutdown.

Whether NRC takes enforcement actions in response to plant performance
problems depends on whether there is a violation of a specific regulatory
requirement.10 Some findings can have risk significance without violating
a regulatory requirement because the ROP is a risk-informed process, while
the underlying regulations are not all risk-informed. For example,
regulatory requirements govern plants' safety-related equipment, which
licensees define when their licenses are granted as equipment that would
be employed to mitigate the effects of an accident. NRC's risk-informed
oversight process considers the condition of all of the equipment at a
plant, not necessarily just that included in its definition of
safety-related equipment. As such, performance deficiencies could be
identified through inspection findings that were not associated with
safety-related equipment and, thus, were not violations of a regulatory
requirement. More specifically, fire pumps are not typically defined as
safety-related equipment under the regulations defining safety-related
equipment, but they could be employed to provide water to cool the reactor
in the event of an accident. If a licensee specified the use of fire pumps
as part of its strategy to mitigate the effects of an accident, a
performance problem associated with the pumps would not necessarily
violate regulatory requirements, but the problem could be determined to be
an inspection finding. Even though there is not a regulation requiring the
licensee to correct such a problem, the finding would be considered for
additional oversight under the ROP, thus providing an incentive for the
licensee to correct it. NRC program officials acknowledge that applying
the risk-informed ROP to regulations that are not risk-informed allows for
these types of situations. The officials said NRC has efforts under way to
risk-inform some of its regulations, which should reduce the likelihood of
these types of situations occurring.

On the basis of the results of its oversight process, NRC provides plant
licensees and the public with an overall assessment of each plant's
performance. At the end of each 6-month period, NRC issues an assessment
letter to each plant to describe what level of oversight the plant will
receive according to its placement on the action matrix, what actions NRC
is expecting the plant licensee to take as a result of the performance
issues identified, any specific enforcement actions NRC has taken, and any
documented substantive cross-cutting issues. If a substantive
cross-cutting issue is identified, the letter will describe what actions
NRC intends to take to monitor the issue and how the licensee is expected
to respond to NRC with the corrective actions it intends to take. NRC also
holds an annual public meeting at or near each site to review its
performance and address questions from members of the public and other
interested stakeholders. In addition, NRC reviews the conclusions of
independent plant assessments, such as those conducted by INPO. The
purpose of this review is to self-assess the NRC inspection and assessment
process to ensure that NRC is identifying similar performance issues.

NRC communicates the results of much of its oversight process to members
of the public through an Internet Web site devoted to the ROP. This Web
site makes available plants' inspection reports and assessment letters,
and other general materials related to NRC's oversight process. NRC also
provides a quarterly summary of every plant's performance, consisting of
its inspection findings, the color of each performance indicator, and its
placement on the action matrix. NRC also provides a short description of
each inspection finding issued during the quarter. While each description
contains information about whether the finding was associated with a
cross-cutting issue, the Web site itself does not provide information on
those plants that have open substantive cross-cutting issues. This
information can only be found on the Web site by linking to each plant's
individual assessment letters. NRC program officials acknowledge that,
without having the information available in summary form, it is difficult
to determine which plants have substantive cross-cutting issues open. They
said that they may look at the possibility of including such information
on the Web site, although doing so is not part of their current plans.

In addition to its plant-level assessments, NRC assesses the results of
its oversight process on an industry-level basis. NRC management holds an
annual meeting to (1) discuss any significant performance issues
identified at specific plants and (2) analyze the overall results of its
inspection and

performance indicator programs and compare them with other
industry-collected and reported performance data. NRC program officials
said that if they identified any negative trends or inconsistencies, they
would take action to better understand and address the cause.

NRC Has Identified Low Risk Problems at Nuclear Power Plants, Resulting in
Increased Oversight for Varying Periods

The ROP has identified thousands of performance deficiencies through
inspection findings at nuclear power plants between 2001 and 2005, but
most of these findings were considered to be of very low risk to safe
plant operations. In the performance indicator program, there have been
very few instances in which performance indicator data were below
acceptable standards during this period. While the majority of plants
received some level of increased oversight due to greater-than-green
inspection findings and performance indicators, only 5 plants were
subjected to NRC's highest levels of oversight. Plants in this category
were generally subjected to this higher oversight for long periods due to
the more intensive supplemental inspections conducted by NRC and to the
more systemic nature of the plants' performance problems and subsequent
corrective actions NRC expected the licensees to take. In assessing the
results of the ROP between 2001 and 2005, we found an association between
poorer performing plants and deficiencies in the cross-cutting areas of
human performance and problem identification and resolution.

Of the more than 4,000 inspection findings identified between 2001 and
2005, 97 percent were green. (See app. III for additional site-specific
inspection findings data.) The number of green findings ranged from 15 at
one site to 141 at another site, with an overall site average of 59 for
the 5-year period. While green findings are considered to be of very low
safety significance, they represent a performance deficiency on the part
of the licensee and it is important that they be corrected. Green findings
consist of such things as finding that a worker failed to wear the proper
radiation detector, or finding that a licensee did not properly evaluate
and approve the storage of flammable materials in the vicinity of
safety-related equipment. NRC does not follow up on the corrective action
taken for every green finding; rather, it relies on the licensee to
address and track the finding's resolution through the plant's corrective
action program. NRC does, however, periodically follow up on some of the
actions taken by the licensee to address green findings through an
inspection specifically designed to evaluate the effectiveness of the
licensee's corrective action program. NRC program officials stated that
even though they do not increase oversight as a result of green findings,
green findings are assessed for the presence of cross-cutting aspects and
provide useful information on plant performance. NRC inspectors told us
they use green findings to identify performance trends in the various
safety areas and these findings help inform their selection of future
inspection samples.

The other 3 percent of the inspection findings identified between 2001 and
2005, or 98 of the more than 4,000, were greater-than-green (see app.
III). Eighty-six findings were white, meaning they were considered to be
of low to moderate risk significance. The white findings were
well-distributed among the sites, with only 11 sites receiving more than 2
white findings during the period. Twenty-three sites were not issued any
white (or greater) findings during this period. White findings were issued
for such things as (1) a licensee's failing to correct a condition in
which the auxiliary service water pump could not be aligned in sufficient
time to mitigate the loss of feedwater during a simulated tornado exercise
and (2) a licensee's having an improper validation process for its
licensed operator requalification examinations, which compromised the
integrity of the exams. Less than 1 percent of the inspection findings
issued since 2001 were of the highest risk significance-yellow or red.
During this time, NRC issued 7 yellow findings and 5 red findings. The 7
yellow findings were in the mitigating systems and emergency preparedness
cornerstones and included issues such as the failure to develop and
maintain emergency plans at one plant, the failure to ensure that
safety-related equipment was adequately protected from flooding at another
plant, and the failure to pass an annual licensed operator requalification
examination by over half of the operator crews at one plant. Of the 5 red
findings, 1 was issued for the degradation of the reactor vessel head at
the Davis-Besse plant, 1 was issued for a steam generator tube failure at
another plant, and the remaining 3 red findings were issued at 2 other
plants for problems related to their auxiliary feedwater pumps.

NRC's inspection results have remained relatively consistent from 2001 to
2005. During this time frame, the number of green findings at all plants
ranged from 657 to 889 per year, and the number of other findings ranged
from 10 to 30 per year, with no strong trend (see fig. 3).

Figure 3: Number of Inspection Findings at All Plants, 2001 Through 2005

Only in the area of inspection findings for which one or more
cross-cutting aspect was associated, is an increasing trend evident (see
fig. 4). According to NRC, the increase in findings with cross-cutting
aspects is due, in part, to the recent development of additional guidance
for inspectors on the identification and documentation of cross-cutting
aspects. The number of plants where the existence of cross-cutting aspects
resulted in NRC's opening a substantive cross-cutting issue is shown in
figure 5. (See app. III for additional data on plant-specific, substantive
cross-cutting issues.)

Figure 4: Number of Inspection Findings and Inspection Findings with
Cross-Cutting Aspects at All Plants, 2001 Through 2005

Note: Almost all inspection findings with cross-cutting aspects were in
the areas of human performance and problem identification and resolution;
only 5 out of 1,890 inspection findings were assigned to the
safety-conscious work environment cross-cutting area.

Figure 5: Number of Plants with Substantive Cross-Cutting Issues, 2001
Through 2005

Note: The totals include those plants that had one or more substantive
cross-cutting issue open during at least some portion of the year. If a
plant had more than one substantive cross-cutting issue open, only one was
counted.

In the performance indicator program, almost all of the indicator data
were reported to be within the acceptable levels of performance (green).
Only 156, or less than 1 percent of over 30,000 indicator reports from
2001 to 2005, did not meet the acceptable performance threshold. (See app.
III for additional plant-specific performance indicator data.) Four of the
15 performance indicators have always been reported within acceptable
performance levels. These 4 indicators include 1 that measures the amount
of time that the residual heat removal safety system is unavailable, 1
that monitors the integrity of a radiation barrier, 1 that measures the
participation of emergency response organizations in emergency drills or
exercises, and 1 that monitors radiological releases. In addition, 46
plants have never had a performance indicator fall outside of the
acceptable performance level, and no yellow or red indicators were
reported during the period.

On the basis of the greater-than-green inspection findings and performance
indicators, NRC has subjected more than three quarters (79) of the 103
operating plants to at least some level of increased oversight for varying
periods. (See app. III for additional plant-specific oversight level
data.)11 Most of these plants received the lowest level of increased
oversight, consisting of the least intensive supplemental inspection, due
to the identification of one or two white inspection findings or
performance indicators. In most of these instances, the licensees did not
accumulate additional greater-than-green findings, so oversight did not
have to be increased to higher levels on the action matrix. For example,
at one plant, NRC issued a white finding after a degraded valve resulted
in the failure of a drain line from a moisture separator. After allowing
time for the licensee to analyze the root cause, NRC completed a
supplemental inspection to assess the licensee's evaluation and concluded
that the licensee's actions were adequate. As a result, NRC did not
require any further actions of the licensee, and NRC inspectors did not
conduct further inspections beyond the baseline procedures. In some cases,
however, either the licensee was not able to sufficiently correct the
performance problem, thus triggering increased oversight, or additional
greater-than-green inspection findings or performance indicators were
identified, also triggering increased oversight. For example, a white
finding was issued at one plant because the licensee's fire response
procedures were not effective in ensuring a safe shutdown of the reactor
during a postulated severe fire. Because this finding was issued during
the same time that one of the plant's performance indicators was white,
the plant was moved from the second to the third highest level of
oversight and NRC conducted a second-level, more-intensive supplemental
inspection. In this case, NRC determined that the licensee's corrective
actions were sufficient to resolve the deficiencies related to both the
inspection finding and performance indicator and thus reduced its
oversight back to baseline inspections only.

While more than three quarters of nuclear power plants received some level
of increased oversight from 2001 through 2005, only 5 plants were placed
into NRC's highest oversight category on the action matrix that still
allows continued operations, and no plants were placed into NRC's

unacceptable performance category.12 Plants subject to the highest level
of NRC oversight generally remain in this category for longer periods of
time than the other performance categories on the action matrix. It
usually takes NRC longer to conduct the more intensive supplemental
inspections and the licensees longer to implement the actions that NRC
expects to correct the broader and more systemic performance issues that
led to their placement in the highest oversight category. For example, one
plant was subject to NRC's highest oversight level for approximately 2
years because of several greater-than-green inspection findings and
performance indicators, including a red inspection finding for the failure
of a steam generator tube. In this case, NRC conducted its most intensive
supplemental inspection 2 months after the final significance of the red
inspection finding was determined. NRC inspectors found a number of
underlying problems with the licensee's programs for design control, human
and equipment performance, problem identification and resolution, and
emergency preparedness. The licensee prepared a plan to address its
performance deficiencies, as is required for all plants placed into this
oversight category, and determined that a multiyear effort was necessary
to adequately develop and implement all corrective actions. Once the
corrective actions were in place, NRC inspectors conducted follow-up
inspections to examine the adequacy of the licensee's efforts in
implementing its corrective actions. NRC inspectors at the plants we
reviewed explained that plants subjected to the highest oversight levels
typically have underlying problems that can take longer periods to
resolve. It is important for them to ensure that the corrective actions
taken by the licensee are effective before oversight is reduced.
Therefore, inspectors try to allow a sufficient amount of time to pass
after the licensee has taken a corrective action, to be able to examine
not only the corrective action itself but also whether it is working as
intended, which requires it to have been in place for awhile.

NRC inspectors at the plants we reviewed also said that a decline in plant
performance is often the result of ineffective licensee corrective action
programs, problems related to human performance, or complacent management,
which often results in deficiencies in one or more of the cross-cutting
areas. In assessing the results of the ROP data, we found an association
between plants that had been subjected to increased levels of oversight
and the presence of substantive cross-cutting issues. For instance, all
plants subjected to NRC's highest level of oversight also had a human
performance and/or problem identification and resolution substantive
cross-cutting issue open either prior to or during the time that they were
subjected to increased oversight.

NRC Is Addressing Weaknesses in Various Areas of Its Oversight Process,
but More Effort Is Needed

NRC has taken a proactive approach to constantly improving its oversight
process over its first 6 years of implementation, but additional
improvements are needed. NRC has several mechanisms in place to
incorporate feedback from both external and internal stakeholders, and it
is currently working on improving several areas of its oversight process
by, for example, better focusing inspections on areas most important to
safety, improving the timeliness and quality of determining the risk
significance of its inspection findings, and modifying some of the
performance indicators to improve their quality. NRC is also assessing
whether it needs to modify its oversight, including developing additional
inspection procedures, as a result of some problems that have surfaced in
areas not fully inspected by NRC, such as the recent discovery of
groundwater contamination from radioactive releases at a number of sites.
In addition, NRC is working to address what we believe has been a
significant shortcoming by making changes to its oversight process to
improve its ability to better identify and address early indications of
deteriorating safety at plants before performance problems develop. Some
of its changes have been controversial, however, and NRC officials
acknowledge the need to carefully assess the effectiveness of the changes.

NRC Is Taking Action to Improve Various Areas of Its Oversight Process

According to NRC, the ROP was implemented with the understanding that it
would be an evolving process and improvements would be made as lessons
were learned. Each year NRC conducts a self-assessment of its oversight
process to compile feedback from various internal and external sources and
to outline the changes it intends to make. As a part of its
self-assessment, NRC developed numerous performance metrics for its
various oversight components, including its inspections and performance
indicator program, to provide quantitative insights into the timeliness,
efficiency, and overall effectiveness of the ROP. The metrics are based on
program data, such as the number and color of inspection findings, and
feedback received from external and internal stakeholders. Each fall, NRC
solicits feedback from external stakeholders through a survey published in
the Federal Register. In 2005, NRC received 21 responses from the
industry, industry organizations, public interest groups, state or local
agencies, and members of the public. NRC also biannually surveys NRC
management, program staff, and regional and site inspectors on the
effectiveness of the ROP. In the most recent survey, which was conducted
in 2004, NRC received responses from about 50 percent of the surveyed
staff. In addition, NRC has a formal feedback mechanism whereby NRC staff
can submit recommendations for improving various oversight components, and
NRC staff meet with industry officials on a monthly basis-in addition to
holding various meetings, workshops, and conferences-to discuss the ROP.
Through NRC's self-assessment process, its staff also incorporates
direction provided by the NRC commissioners and recommendations from
independent evaluations, such as from the ACRS, GAO, and the NRC Office of
the Inspector General.

According to NRC program officials, the changes made to the ROP since its
inception in 2000-including those made in response to the Davis-Besse
incident-have generally refined the existing oversight process, rather
than significantly changing it. In the case of Davis-Besse, NRC formed a
task force to review the agency's regulatory processes. The task force's
report, issued in September 2002, contained more than 50 recommendations,
many associated with its oversight process. Among the more significant
ROP-related recommendations were those to (1) enhance the performance
indicator that monitors unidentified leakage; (2) develop specific
guidance to inspect licensee boric acid control programs and vessel head
penetration nozzles; (3) modify the inspection program, including
modifying an inspection procedure to better follow up on long-standing
plant performance issues; and (4) enhance the guidance for managing plants
that are shut down as a result of significant performance problems. NRC
program officials told us that the task force's most significant
recommendations were in areas outside of the ROP, such as improving the
agency's operating experience program. According to NRC, it has
implemented almost all of the task force's recommendations.

Other modifications NRC has recently made or is in the process of making
include the following:

o NRC recently revised 7 of its baseline inspection procedures to better
focus the level and scope of its inspection efforts on those areas most
important to safety. These revisions resulted from a detailed analysis in
2005 of its more than 30 baseline inspection procedures. For example, NRC
staff analyzed the number of findings resulting from each of its
inspection procedures to better understand the areas where performance
deficiencies were occurring, and compared the time that inspectors were
spending directly observing plant activities as opposed to reviewing
licensee paperwork to achieve a balance between the two. For this effort,
NRC held constant the level of resources devoted to its baseline
inspection effort and did not assess whether more or fewer inspection
resources overall should be applied. NRC is now formalizing this analysis
and making it a regular part of its yearly assessment so that it can
continually refine its inspections to direct resources to the most
critical performance areas.

o NRC has efforts under way to improve its SDP. An audit by the NRC Office
of the Inspector General, a review by a special task group formed by NRC,
and feedback from other stakeholders such as NEI and UCS have pointed to
several significant weaknesses with the SDP. Also, other internal and
external stakeholders, including NRC inspectors and plant managers, have
raised concerns about the amount of time, level of effort, and knowledge
and resources required to determine the risk significance of some
inspection findings. Industry officials commented that because most
inspection findings are green, one white finding at a plant can place it
in the "bottom quartile" of plants from a performance perspective.
Therefore, industry officials explained, licensees try to avoid this
placement and will expend a great deal of effort and resources to provide
additional data to NRC to ensure that the risk level issued for the
finding is appropriate. This can add significant time to the process
because licensees may use their own technical tools, including models,
data, and assumptions, to analyze the issue. NRC then considers this
information in its own analysis. At the plants we reviewed, the time it
took NRC to determine the final significance of risk-significant
inspection findings ranged from 24 days to over 200 days. The delay in
assigning a color to a finding while the new information is being
considered could affect a plant's placement on NRC's action matrix,
essentially delaying the increased oversight called for if the finding is
determined to be greater-than-green. NRC program officials said that even
though there may be a delay in determining the final risk significance of
the finding, the licensee takes immediate action to fix the identified
problem. However, supplemental inspections are designed to uncover
additional associated problems, if they exist, and delaying these
inspections could delay NRC's ability to discover additional performance
problems, potentially allowing performance to worsen.

NRC developed a plan in 2002 to improve the timeliness of its
decision-making process, enhance the quality of its various SDP tools, and
track its progress in implementing key changes. For example, NRC
introduced a new process aimed at improving timeliness by engaging
decision makers earlier to more quickly identify the scope of the
evaluation to be used to determine the inspection finding's risk
significance, the resources needed, and the schedule to complete the
evaluation.

As a part of the SDP, NRC uses probabilistic risk assessment (PRA)
methods, which overall has improved its ability to assess the performance
and safety of nuclear power plants. PRA is an analytical tool for
estimating the probability that a potential accident might occur by
examining how physical structures, systems, and components, along with
employees, work together to ensure plant safety. Using PRA tools, NRC and
the plant licensees can estimate the likelihood that different accident
scenarios at plants will result in reactor core damage and a release of
radioactive materials. NRC often uses PRA tools to help it determine the
risk significance of its inspection findings. However, we and others have
found weaknesses with NRC's use of PRA.13 For instance, in our May 2004
report on the Davis-Besse reactor vessel head incident, we found that NRC
used some incomplete and faulty PRA analyses in deciding whether to allow
the licensee to delay shutdown of the reactor for inspection.14 While NRC
program officials acknowledge they can improve their current PRA tools,
they said the tools are adequate for factoring risk into the oversight
process and do achieve their intended purpose. NRC has several initiatives
under way to improve its use of PRA in its decision making, many of which
it plans to complete by June 2007. For example, NRC recently revised its
models to assess events that occur at plants during at-power situations to
better capture individual plant characteristics, and it is in the process
of developing new models to assess events that occur at plants during
low-power or shutdown conditions. NRC is also developing guidance to
address broader PRA issues, including establishing overall quality
requirements for risk information; providing specific instructions for
documenting the decision process and conclusions; and developing improved
methods for calculating risk, such as its practices for implementing human
reliability analysis.

o NRC is working with the nuclear power industry to improve several of its
performance indicators, including making the definitions of some
indicators more concise to improve their quality and to reduce the number
of discrepancies between licensees and NRC inspectors. Because NRC
inspectors verify indicator data only once a year, a potential
disagreement with a licensee over these data might not surface for up to 1
year after it is reported, and it may take even longer to resolve the
disagreement. Similar to delays with the SDP, a delay in assigning a color
while the disagreement is being resolved could affect a plant's placement
on NRC's action matrix, and delay the increased oversight called for if
the indicator is determined to be greater-than-green. To date, NRC's
efforts to improve the performance indicator program have largely centered
on developing a key indicator to address known problems with the
indicators that measure the unavailability of safety systems. NRC has been
working jointly with the nuclear industry since 2001 to develop the
indicator, which is now risk-informed and intended to provide a more
accurate indication of the risks associated with changes in the
availability and unreliability of important safety systems. This
indicator, referred to as the mitigating systems performance index, was
implemented in April 2006, and the first plant reports were submitted in
July 2006. NRC is also in the process of changing the definition for
several other indicators, in addition to considering the feasibility of
new indicators.

In addition, NRC program officials said they are taking a broad look at
the performance indicator program to assess how they might increase its
overall value to the ROP. The effort is intended, in part, to address
concerns that performance indicators have not contributed to the early
identification of poorly performing plants to the degree originally
envisioned, since all of the indicators are almost always within
acceptable performance levels (green). When the program was developed,
thresholds were set using industry performance data such that about 5
percent of the plants were expected to exceed the green-white
threshold-that is, be designated either white, yellow, or red-for each of
the performance indicators. However, from 2000 through 2005, less than 1
percent exceeded it. Furthermore, there have been several cases where
plants reported an acceptable performance indicator and performance
problems in the area were subsequently identified. For example, NRC
inspectors at one plant noted that while performance indicator data
related to its alert and notification system for emergency preparedness
had always been reported green, the system did not always function
properly. On the other hand, industry officials believe that the high
percentage of green indicators is indicative of plants' good performance.
Several plant managers told us that they closely monitor and manage to the
acceptable performance thresholds established for each indicator, and that
they will often take action to address performance issues well before the
indicator crosses the acceptable threshold. To assess issues within the
performance indicator program, NRC developed a standing working group,
with representatives from both the industry and NRC, that meets on a
monthly basis. This group, along with senior NRC and industry management,
plans to explore overall program improvements to better identify declining
plant performance.

In addition to the previously mentioned efforts, NRC is assessing whether
it needs to modify its oversight, including developing additional
inspection procedures, as a result of some problems that have surfaced in
areas not fully inspected by NRC. For example, NRC has reported that there
have been inadvertent, unmonitored releases of radioactive liquids
containing tritium at a number of nuclear power sites in the past few
years.15 To date, NRC reports show that the measured levels of tritium
discovered were low enough that they do not appear to pose a public health
hazard. According to NRC, the releases were due to equipment failures or
structural degradation at the plants. At one of the plants we reviewed,
for example, elevated levels of tritium in the on-site groundwater were
discovered during the licensee's testing of its monitoring wells. NRC
reports suggest that this radioactive release was most likely due to
leakage from the spent fuel pool's support structures. Shortly after the
licensee reported the contamination, NRC initiated a special inspection to
further investigate the source and cause of the leakage and the licensee's
actions for mitigation. On the basis of this and other discoveries of
contaminated groundwater, in March 2006, NRC formed a lessons-learned task
force to evaluate NRC's regulatory processes related to the radioactive
releases and to recommend areas for improvement. For example, NRC is
reviewing related regulations and guidance, inspection program
requirements, and its communications with external stakeholders and the
public. In addition, the nuclear industry recently undertook an initiative
to improve its groundwater protection, and intends to develop
site-specific action plans at all plants and improve the industry's data
collection, reporting, and protocols for sharing lessons learned.

NRC is also taking action to improve its requirements for licensees to
control and account for their spent nuclear fuel, that is, the used fuel
periodically removed from reactors in nuclear power plants. NRC requires
licensees to control and account for all of their spent fuel materials
because of the hazardous nature of spent nuclear fuel.16 However, reviews
by both GAO and the NRC Office of the Inspector General found weaknesses
with NRC's oversight of licensees' spent fuel control and tracking
programs.17 Between 2000 and 2004, several plants experienced instances of
missing or unaccounted-for spent fuel, and NRC reported weaknesses in the
material control and accounting programs at various other plants. In
investigating these issues, we and the NRC Office of the Inspector General
determined that NRC's inspections did not adequately ensure that all
licensees properly controlled and accounted for their spent nuclear fuel.
From the late-1980s through implementation of the ROP, NRC did not conduct
routine material control and accounting inspections; instead it looked at
these activities indirectly through other inspections, such as those of
licensee operations during refueling of the reactor. We also found that
while NRC requires plants to maintain an accurate record of all their
spent fuel and its location, NRC regulations did not specify how licensees
are to conduct physical inventories or how they are to control and account
for loose spent fuel rods and fragments. NRC is currently in the process
of revising several of its guidance documents and developing inspection
procedures to better assess the effectiveness of licensee material control
and accounting programs. NRC reports that most of these efforts will be
completed by the end of 2007.

NRC Is Taking Its First Major Step to Address a Significant ROP Weakness
in the Area of Safety Culture

One significant shortcoming in the ROP that we and others have found is
that it has not been as effective as it could be in identifying and
addressing early indications of deteriorating safety performance at
nuclear power plants before problems develop. NRC and others have long
recognized that a safety culture-the organizational characteristics that
ensure that issues affecting nuclear plant safety receive the attention
their significance warrants-can have a significant impact on a plant's
safety performance. The identification of a weak safety culture-or
weaknesses in the attributes that make up a safety culture, such as
attention to detail, adherence to procedures, and effective corrective and
preventative actions-can point to early signs of deteriorating safety
performance before conditions become so serious that a safety accident
occurs. As early as 1989, NRC recognized the importance of developing a
safety culture at each nuclear power plant that ensures safe plant
operations, but NRC's policy stated that it was the licensee's duty and
obligation to monitor and maintain a strong safety culture.

Despite the recognition of the importance of a safety culture and several
external groups' recommendations to better incorporate safety culture
aspects into NRC's oversight process, NRC did not include specific
measures to comprehensively assess plant safety culture when it
implemented the ROP in 2000. As its new oversight process was being
developed, external stakeholders, including the ACRS, concluded that
additional oversight measures were needed to characterize licensees' human
performance and safety culture. The 2002 Davis-Besse reactor vessel head
incident highlighted that this was a significant weakness in the ROP. Our
May 2004 report concluded that the event occurred, in part, because NRC
did not have an effective means to identify and address early indications
of deteriorating safety at plants before performance problems develop.

NRC did not take immediate action, however, contending that direct safety
culture evaluations would cross the line from a regulatory function to a
licensee management function. In August 2004, the NRC Commission directed
the staff to improve the ROP by more fully addressing safety culture. In
response, NRC staff formed a safety culture working group in early 2005 to
lead the agency's efforts to make changes to the ROP to better incorporate
safety culture into its oversight process. As a part of this initiative,
the working group obtained the input of external stakeholders through a
series of public meetings held in late 2005 and early 2006. The group also
incorporated lessons learned from events that occurred at the Salem and
Hope Creek site during this same period. In 2004, NRC confirmed there were
problems with aspects of the site's safety culture. In addressing the
problem, NRC concluded that the ROP did not provide adequate tools to
monitor the situation or the sufficiency of the licensee's corrective
actions. As a result, NRC deviated from the ROP to increase its oversight
at the site, which included conducting special inspections and forming a
team with expertise in the area to review the licensee's corrective
actions. (See app. I for additional information on the events at Salem and
Hope Creek.)

In February 2006, NRC issued proposed changes to some of its inspection
procedures and guidance documents to incorporate safety culture into the
ROP, and implemented the changes in July 2006. NRC used the following two
overall approaches: first, it developed additional guidance for
identifying and addressing cross-cutting aspects, and, second, it
developed a structured way to determine the need to evaluate plants'
safety culture. Several inspection procedures were also modified to direct
inspectors to be sensitive to and take into consideration safety culture
components when planning and conducting their inspections. Although the
three cross-cutting aspects (problem identification and resolution, human
performance, and a safety conscious work environment) did not change, NRC
developed new definitions for them to more fully encompass safety culture
aspects. Also, NRC developed additional guidance on the treatment of
cross-cutting aspects once they have been identified. For example, the
problem identification and resolution cross-cutting area now has several
components-a corrective action program, operating experience, and self and
independent assessments. The human performance cross-cutting area is
composed of decision making, resources, work control, and work practices.
NRC inspectors are required to assess every inspection finding to
determine if it is associated with one or more of the components that make
up each of the cross-cutting areas. While the process for assessing
inspection findings for the existence of cross-cutting aspects and
substantive cross-cutting issues remains largely unchanged, now the
definitions for the cross-cutting areas are more detailed and track more
closely with those elements that comprise safety culture.18 In addition,
under NRC's new guidance, if the same substantive cross-cutting issue is
identified in three consecutive assessment periods, NRC may request that
the licensee evaluate its safety culture. The intent is to provide an
opportunity to diagnose a potentially declining safety culture before
significant safety performance problems occur. NRC program officials said
they consider the identification and treatment of substantive
cross-cutting issues the most proactive element of the ROP, because all
other oversight actions are taken only when more significant performance
problems have been identified.

NRC's changes to the ROP now also include a structured way for NRC to
determine the need for a safety culture evaluation. NRC's new guidance
calls for the licensees of plants with more than one white finding in the
same cornerstone or one yellow finding to evaluate whether the performance
issues were caused by any safety culture components, and NRC may request
the licensee to independently evaluate its safety culture, if the licensee
does not identify a safety culture component. Any safety culture
deficiencies are expected to be entered into the licensee's corrective
action program. Regional officials would discuss the licensee's proposed
corrective actions with the licensee, and NRC may hold a public meeting to
discuss the issues. For plants where more significant or multiple findings
have been identified, NRC will not only independently assess the adequacy
of the licensee's independent evaluation of safety culture, but may also
conduct its own evaluation. Following the completion of any evaluations,
regional and headquarters officials together determine whether additional
agency actions are warranted, and, at a minimum, the licensee will be
required to document its plan to make improvements and a public meeting
will be held to discuss the licensee's performance. According to an NRC
official familiar with the Salem and Hope Creek situation, had these
requirements been in place at the time, they would have been adequate to
address the concerns at Salem and Hope Creek without the need to deviate
from the ROP.

NRC's approach to incorporating safety culture into the ROP has been
controversial, and some stakeholders disagree with certain changes. For
example, some in the nuclear power industry have expressed concern that
the changes could introduce undue subjectivity to NRC's oversight, given
the difficulty in measuring these often intangible and complex concepts.
Several of the nuclear power plant managers at the sites we reviewed said
that it is not always clear why a cross-cutting aspect is associated with
a finding, or what it will take to clear themselves once they have been
identified as having a substantive cross-cutting issue. Some industry
officials worry that the changes will further increase the number of
findings that have cross-cutting aspects associated with them, and, if all
of the findings have these aspects, whether the process will lose its
value. Industry officials also warn that if the changes are not
implemented carefully, resources could be diverted away from other
important safety issues. Other external stakeholders, such as an official
from UCS, on the other hand, suggest that this effort is an important step
toward improving NRC's ability to identify performance issues at plants
before they result in safety problems. Importantly, there are now
additional tools in place for NRC to use when it identifies potential
safety culture concerns. In reviewing NRC's proposed approach in April
2006, the ACRS concluded that the approach was appropriate and will
enhance the agency's ability to address safety culture issues; although
after gaining some experience with the process, it stated that NRC should
reassess the adequacy of some procedures.

NRC program officials acknowledged that they will need to assess the
changes they made to the ROP to determine if they better allow inspectors
to detect deteriorating safety conditions at plants before significant
safety events occur. Some at NRC view these changes as the beginning step
toward an incremental approach. For its current efforts, NRC also
acknowledged that additional training for its inspectors on safety culture
is needed. NRC provided computer-based and regional training to its
inspectors as the changes were being finalized, and it is currently
working to incorporate aspects of its safety culture changes into its more
permanent training programs. NRC plans to evaluate stakeholder
feedback-both through its normal processes, such as its monthly meetings,
and potentially through specific industry-sponsored workshops dedicated to
the issue-and make additional changes on the basis of the lessons learned
as part of its annual self-assessment process for 2007.

Conclusions

NRC is devoting considerable effort to overseeing the safe operation of
the nation's commercial nuclear power plants, and its process for doing so
appears logical and well-structured. NRC's oversight process is finding
safety problems and is getting the industry to constantly improve.
However, weaknesses with its inspection and performance indicator programs
have been identified-in particular, the timeliness of the process used to
determine the risk significance of inspection findings, and the ability of
performance indicators to contribute to the early identification of poorly
performing plants. Importantly, NRC is demonstrating that it is aware of
these weaknesses and is actively making changes to improve its oversight.
NRC's proactive approach is demonstrated by the important progress it is
making in several key areas, including making efforts to improve the
timeliness and quality of its significance determination process,
redefining some of its performance indicators, and assessing the need for
additional inspection procedures based on careful analysis. Its efforts to
continuously consider the need to improve and obtain feedback from both
internal and external stakeholders are critical as nuclear power plants
age and the nation considers building new plants. In this regard, it is
also important that the ROP continue to be a very open process in which
NRC provides the public and its other stakeholders with considerable
information on its oversight activities and findings related to plant
safety performance.

Although NRC has been working to improve its oversight in several key
areas, its efforts to incorporate safety culture into the ROP may be its
most critical future change. More than 4 years have passed since
Davis-Besse highlighted that a significant weakness in NRC's oversight was
its inability to identify deteriorating safety conditions at plants before
they resulted in performance problems. NRC has been reluctant to regulate
in the area of safety culture because it did not want to be directly
involved in managing the licensees' plants. However, NRC is now taking
concrete actions to begin incorporating safety culture into the ROP,
although it acknowledges that regulating the often complex and intangible
aspects of safety culture is challenging, and that its recent changes are
simply a first step. As a result, it will be important to closely monitor
this effort to ensure that it is achieving the goal of objectively
assessing safety culture, while providing an early indication of declining
safety performance. An additional challenge for NRC will be how to provide
information to the public and other stakeholders on this important but
complex area of plant performance. Given that it may take some time for
NRC to develop performance metrics for safety culture, data on substantive
cross-cutting issues, which provide insight into aspects of plants' safety
culture, could be useful to the public and other stakeholders as they look
for assurances that plants are operated safely. Summary-level information
on plants with substantive cross-cutting issues is not currently available
to the public through NRC's Web site.

Recommendations for Executive Action

Given its importance to improving NRC's ability to identify declining
safety performance at nuclear power plants before significant safety
problems develop, we recommend that the NRC commissioners take the
following two actions:

o Aggressively monitor; evaluate; and, if needed, implement additional
methods or processes to increase the effectiveness of its efforts under
the ROP to assess safety culture at plants.

o In addition to periodically evaluating the effectiveness of its safety
culture efforts, NRC may also be able, through its performance indicator
program, to develop specific indicators to measure important aspects of
plants' safety culture. Trends in these performance indicators could be
useful feedback to NRC on its safety culture activities. The indicators
could also provide useful information to the public and other NRC
stakeholders on the safety culture at plants.

In addition, in the absence of performance indicators or other performance
metrics for plants' safety culture, we recommend that the NRC
commissioners make publicly available, through the ROP Web site,
consolidated and comprehensive data on the plants that have substantive
cross-cutting issues open. These data would provide a more comprehensive
picture of plant performance and provide insights into aspects of the
plants' safety culture that otherwise are not readily available on the Web
site.

Agency Comments

We provided a draft of this report to NRC for its review and comment. In a
letter from NRC's Executive Director for Operations, NRC generally agreed
with the report's findings, conclusions, and recommendations (see app.
IV). NRC also commented that the report is comprehensive, fair, and
balanced. In addition, NRC provided minor, technical comments, which we
have incorporated into the report, as appropriate.

We are also sending copies of this report to interested congressional
committees, the Chairman of NRC, and other interested parties. We also
will make copies available to others upon request. In addition, the report
will be available at no charge on the GAO Web site at http://www.gao.gov .

If you or your staffs have any questions about this report, please contact
me at (202) 512-3841 or [email protected] . Contact points for our Offices of
Congressional Relations and Public Affairs may be found on the last page
of this report. GAO staff who made major contributions to this report are
listed in appendix V.

Jim Wells Director, Natural Resources and Environment

List of Requesters

The Honorable James M. Inhofe Chairman Committee on Environment and Public
Works United States Senate

The Honorable George V. Voinovich Chairman Subcommittee on Clean Air,
Climate Change,   and Nuclear Safety Committee on Environment and Public
Works United States Senate

The Honorable Ed Whitfield Chairman Subcommittee on Oversight and
Investigations Committee on Energy and Commerce House of Representatives

The Honorable Joseph R. Biden, Jr. United States Senate

The Honorable Tom Carper United States Senate

The Honorable Michael N. Castle House of Representatives

The Honorable Edward J. Markey House of Representatives

Appendix I

Key Safety-Related Events at the Salem and Hope Creek Nuclear Power Plants
from 2000 to 2006 

This appendix summarizes key safety-related events at the Salem and Hope
Creek nuclear power plants and the Nuclear Regulatory Commission's (NRC)
and licensee's actions to follow up on and correct them. This information
is presented to enhance the public's understanding of the events and
timing of NRC's actions. We did not conduct an independent audit or
assessment of the events, other than to help illustrate how NRC is using
its oversight process to ensure plant safety.

The Salem and Hope Creek plants are located at one site in Hancocks
Bridge, New Jersey, 18 miles southeast of Wilmington, Delaware, on the
Delaware River. The two Salem plants (Salem 1 and Salem 2) consist of two
pressurized water reactors, and the Hope Creek plant is a boiling water
reactor. The three plants produce a combined 3,237 megawatts of
electricity. Salem 1 began operating in the late-1970s and Salem 2 and
Hope Creek began operating in the 1980s. PSEG Nuclear LLC is licensed to
operate all three plants, with their current licenses expiring roughly 40
years after the start of their operations. In January 2005, PSEG entered
into a nuclear operating services contract with Exelon, the first step of
a planned merger between the two companies. Under the terms of this
contract, Exelon manages operations at Salem and Hope Creek.

From 2000 through 2006, there were several safety-related events at the
site that affected the plants' performance. As a result, NRC increased its
oversight to include supplemental and special inspections, and required
the licensee to take specific actions to address its performance issues.
NRC also deviated from its normal oversight process from August 2004
through 2006 to increase its oversight and address problems with the
licensee's ability to provide an adequate safety-conscious work
environment (SCWE). A SCWE is defined by NRC as an environment in which
employees feel free to raise safety concerns, both to their management and
to NRC, without fear of retaliation. During this period, NRC and the
licensee held several public meetings to discuss performance issues
related to the site's safety work environment, or the SCWE, and the
actions the licensee was taking to improve it.

The following summary provides details on the key safety-related events at
Salem and Hope Creek, along with NRC's and the licensee's responses to the
events, from the initial implementation of NRC's reactor oversight process
in April 2000 to June 2006.

Summary of Key Safety-Related Events at Salem and Hope Creek, April 2000
to June 2006

White Performance Indicator for Unplanned Scrams

During the fourth quarter of 2000, Salem 1 reported a white color for the
unplanned scrams performance indicator, an indicator that monitors the
number of unplanned scrams-reactor shutdowns-that occurred during the
previous four quarters. During 2000, Salem 1 had four scrams, which is one
more than the acceptable performance level. NRC conducted a supplemental
inspection in March 2001 to follow up on why such a high number of scrams
occurred and on the corrective actions the licensee was taking to address
the problem. In conducting its inspection, NRC determined that the
licensee had performed a comprehensive common cause analysis of the
associated performance deficiencies, which identified human performance,
equipment failure, and procedure and preventive maintenance program issues
as root causes of the performance problems. The licensee documented these
issues in its corrective action program and developed a plan to outline
the actions it would take to address them. NRC concluded this effort was
sufficient and did not require the licensee to take any additional
corrective actions beyond what it had outlined in its plan.

White Performance Indicator for Unplanned Changes in Reactor Power

During the first quarter of 2002, Salem 1 reported a white color for the
performance indicator that monitors unplanned changes in reactor power
that could have challenged safety functions during the previous four
quarters. Salem reported a total of seven power changes, which exceeded
the acceptable performance threshold by one. NRC conducted a supplemental
inspection in October 2002 and concluded that the licensee's
implementation of several programs to address the underlying causes of the
unplanned power changes, such as the replacement of certain pieces of
equipment, sufficiently addressed the performance problem. NRC did not
require the licensee to take any additional corrective actions.

Substantive Cross-Cutting Issue in Problem Identification and Resolution

In March 2003, NRC opened a substantive cross-cutting issue in the problem
identification and resolution area at all three plants on the basis of the
identification of a number of green inspection findings that documented
ineffective problem evaluations and untimely, ineffective corrective
actions by plant employees, including recurring equipment failures. Upon
opening this substantive cross-cutting issue, NRC stated that it would
closely monitor the licensee's performance in this area. This substantive
cross-cutting issue remained open until March 2006, when NRC concluded
that the licensee's efforts to improve its ability to identify and resolve
performance problems were sufficient.

White Inspection Finding for Failure of a Part on an Emergency Diesel
Generator

In May 2003, NRC determined that the failure of a part on an emergency
diesel generator at the Salem 1 plant that occurred in September 2002
would result in a white inspection finding. NRC conducted a supplemental
inspection in October 2003 and concluded that although the licensee had
adequately fixed the emergency diesel generator, it would need to take
additional action to address broader performance problems. Specifically,
NRC had concerns with the licensee's ability to ensure that the controls
and procedural requirements determined to be necessary on the basis of
evaluations of equipment failures, were reliably tracked and implemented.
The licensee implemented procedural changes to more effectively track its
corrective actions to prevent recurrence. NRC conducted a second,
follow-up supplemental inspection in September 2004 and, on the basis of
this inspection, concluded that the licensee's corrective actions were
sufficient.

White Inspection Finding for the Failure of a Rotating Screen That Is Part
of the Station Service Water System at Hope Creek

In May 2004, NRC determined that the failure in July 2003 of a rotating
screen that is part of the station service water system at Hope Creek
would result in a white inspection finding. NRC determined that the
failure resulted from the licensee's inadequate maintenance procedures and
its failure to adhere to procedural instructions. In September 2004, NRC
conducted a supplemental inspection to follow up on the corrective actions
that licensee had taken. It determined that the licensee's corrective
actions, which included revising relevant maintenance procedures, were
sufficient and did not require additional actions from the licensee.

White Inspection Finding for the Failure of a Drain Line in the Moisture
Separator System at Hope Creek

In February 2005, NRC determined that the licensee's failure to properly
evaluate and correct a degraded valve, which resulted in the failure of a
drain system at Hope Creek in October 2004, would result in a white
inspection finding. NRC conducted a supplemental inspection in June 2005
to follow up on the licensee's corrective actions, which included the
development of new guidance; revisions to relevant plant operating
procedures; inspections of pipe hangers, one of which was determined to be
the initiator of the degraded valve; and the development of a new
procedure for additional monitoring of degraded equipment. NRC determined
that these actions were sufficient and did not require additional actions
from the licensee.

Increased Oversight to Address Safety-Conscious Work Environment Problems

In late 2003, NRC initiated a special review of the site's SCWE on the
basis of allegations concerning its SCWE, plant events and inspection
findings that indicated problems with its SCWE, insights from NRC
interactions with the licensee, observations made by NRC inspectors, and
the presence of a substantive cross-cutting issue in problem
identification and resolution. NRC's special review consisted of (1)
in-depth interviews by NRC experts of more than 60 current and former
plant employees, (2) an analysis of the site's inspection and assessment
record over the previous several years, and (3) information from plant
employees' allegations related to the SCWE. NRC provided the interim
results of its review to the licensee in January 2004. Although NRC did
not identify any serious safety violations, the information led to
concerns about the site's SCWE, particularly as it related to the handling
of emergent equipment issues and associated operational decision making.
The review accumulated information about a number of events that, to
varying degrees, called into question the openness of management to
concerns and alternative views, strength of communications, and
effectiveness of the licensee's corrective action and feedback processes.
There were several differences of opinion among operators and senior
managers on plant operating decisions, particularly as they might impact
continuing plant operation and outage schedules. On the basis of these
interim results, NRC requested that the licensee conduct an in-depth
assessment of its SCWE.

NRC issued the final results of its special review in July 2004,
confirming many of the concerns it identified through its interim review.
In its final results, NRC concluded that there were weaknesses in the
licensee's leadership and management approaches, leading to a perception
among some staff and mangers that the company emphasized production over
safety. NRC also determined that licensee management was not consistent in
its support of staff identifying concerns and providing alternate views,
and cited examples of unresolved conflicts and poor communication between
management and staff.

In May 2004, the licensee submitted an independent assessment of its SCWE.
The assessment included interviews of employees, and reviews of the
licensee's inspection record and employee concerns program, among other
things. Among the findings of the assessment were that (1) some plant
employees were hesitant to raise issues; (2) management was not receptive
to or effective at addressing some employee concerns, such as those
surrounding long-standing equipment problems; (3) a significant number of
employees did not view the employee concerns program as a viable means to
raise concerns; and (4) management was not effective at understanding or
addressing the potential for a "chilling" effect--that is, an environment
that discourages workers from raising safety concerns--in response to
highly visible employee concerns and actions associated with operational
events.

As a result of the assessment, the licensee submitted to NRC in June 2004
an action plan to improve its overall safety work environment. The plan
addressed the licensee's corrective action program, work management
program, and safety-conscious work environment. For example, to improve
the licensee's corrective action program, the plan identified actions to
improve monitoring, such as developing and implementing an integrated
corrective action training program and developing performance indicators.
The plan also included a number of actions to improve the licensee's
management alignment, prioritization, support for and awareness of the
workweek schedule, and communication and training strategies to support
work management improvements. In addition, the plan included actions to
improve the willingness of plant employees to raise concerns, improve the
effectiveness of policies and procedures for resolution of issues, enhance
key elements of the employee concerns program, and improve management
effectiveness in detecting and preventing retaliation or a chilled
environment. The licensee's action plan included 17 metrics designed to
measure safety work environment improvements. The licensee provided the
results of these metrics to NRC on a quarterly basis.

In August 2004, NRC regional officials received approval from headquarters
to deviate from the normal oversight process to increase oversight at the
site to monitor the licensee's actions to improve its safety work
environment. NRC's increased oversight consisted of (1) reviewing the
results of the specific actions the licensee took as a part of its action
plan, (2) creating an NRC team with expertise in the area to assist with
coordinating and focusing these review efforts, and (3) conducting
additional special inspections and enhancing existing inspections by
adjusting their focus and scope. NRC determined that it could reduce its
oversight to baseline levels once the licensee completed a self-assessment
of its SCWE that concluded that it had made substantial, sustainable
progress, and NRC confirmed the licensee's conclusion.

NRC also opened a substantive cross-cutting issue in the SCWE area at all
three plants in August 2004, on the basis of the results of its special
review of the site's SCWE completed that July. Upon opening this
substantive cross-cutting issue, NRC requested that the licensee discuss
its progress in improving its SCWE and the effectiveness of its corrective
action program in a public meeting planned for late 2004. This substantive
cross-cutting issue remained open in the 2005 annual assessment letter,
issued in March 2006. NRC will consider closing this substantive
cross-cutting issue after the licensee provides the results of an
assessment that concludes that it has made substantial, sustainable
progress, and NRC has completed a review that confirms these results.

In June 2005, NRC issued the results of an inspection of the licensee's
employee concerns program (ECP) conducted as a part of its increased
oversight to monitor the site's safety work environment. NRC inspectors
did not identify any findings of safety significance and concluded that
the ECP provided a framework for investigating concerns, maintaining the
confidentiality of personnel who use the program, and protecting employees
who use the program against retaliation. They also concluded, however,
that it was too early to fully assess the effectiveness of recent program
improvements and initiatives. The inspectors observed that a statistically
significant portion of the personnel interviewed indicated that they would
not use the ECP due to a perception that the process did not take adequate
measures to protect users' confidentiality. NRC inspectors cited a section
of the ECP that appeared to affirm this perception. The licensee
emphasized that it was making efforts to protect confidentiality, and
acknowledged that the ECP should reflect these efforts.

NRC also provided the preliminary results of its review of the licensee's
executive review board (ERB) process in June 2005. The ERB was established
to improve the site's safety work environment and management's
effectiveness in detecting and preventing retaliation and a chilled work
environment. NRC inspectors determined that lapses in the licensee's use
of the ERB process constituted a green inspection finding, although it did
not represent a violation of regulatory requirements. NRC requested that
the licensee reassess the review of the ERB in the broader context of the
work environment, identify additional actions planned or taken to address
negative worker perceptions, and provide a written response to NRC within
30 days. The licensee responded with planned corrective actions, including
developing and implementing continuing training on the SCWE, and
developing and implementing a plan to improve its corrective action
program.

In July 2005, NRC headquarters extended regional officials' permission to
deviate from the normal oversight process and provide increased oversight
at Salem and Hope Creek. The extension was necessary because the licensee
had not yet met the criteria to move back to normal, or baseline levels of
oversight. Increased oversight included (1) continued management meetings
and site visits to review the implementation of the licensee's corrective
actions; (2) increased efforts, including using more inspectors and
samples than what would typically be used, to conduct a baseline
inspection of the site's effectiveness at identifying and resolving
problems; and (3) an additional inspection to monitor the licensee's
progress in resolving the substantive cross-cutting issue in the SCWE.

In November 2005, NRC issued the results of the special inspection on the
site's SCWE. The inspection included assessing the licensee's progress and
plans for making improvements to its SCWE, its metrics to monitor the
effectiveness of the improvements made, and effectiveness of the
licensee's corrective actions and self assessment initiatives. NRC
inspectors did not identify any findings of safety significance and
determined that the licensee had made progress in improving its SCWE. For
example, NRC inspectors concluded that workers' willingness to raise
safety concerns had increased. In addition, the licensee had taken a
significant number of actions to improve the corrective action and work
management programs and had implemented several corrective actions for the
employee concerns program. They also concluded that the licensee had made
progress in preventing and detecting retaliation. However, NRC inspectors
also observed that some issues required additional action and focused
attention, such as the need for the licensee to fully evaluate and address
negative perceptions about its work environment in certain work groups.
The licensee initiated actions to address these observations.

In May 2006, the licensee submitted to NRC an independent peer assessment
that concluded that it had made substantial improvements to its SCWE, and
a solid foundation existed to sustain them. The assessment included
interviews with site personnel; observations of station activities and
meetings; and reviews of the licensee's programs, procedures, policies,
and other relevant information. Among the assessment's conclusions were
that personnel throughout the organization exhibited a willingness to
engage in open-and-candid discussions and raise safety and quality issues,
the corrective action and work management programs had improved,
management had been effective at detecting and preventing retaliation and
addressing chilling effects in response to the raising of safety concerns,
and management provided high-visibility and strong and continuous
reinforcement of good work environment principles. In response, NRC began
a review of the effectiveness of the licensee's actions to improve its
safety work environment. The results of this review will be a key input
into NRC's midcycle assessment of the site's overall safety performance to
be completed in August 2006; through the midcycle assessment NRC will
determine any changes in the level of oversight for Salem and Hope Creek.

Appendix II

Scope and Methodology

To examine how NRC oversees plants, we reviewed the various tools and
processes that comprise the Reactor Oversight Process (ROP). In this
regard, we analyzed NRC's documentation of its oversight process,
conducted interviews with NRC program staff and other officials, visited
one NRC regional office, and visited one nuclear power plant site that was
of specific interest to our requesters. In particular, we reviewed NRC's
policies, inspection manuals, and other guidance documents outlining its
various oversight process components-including reports discussing the ROP
design basis; inspection and reporting requirements; performance indicator
program guidance; and other requirements, such as those related to its
enforcement and assessment processes. We interviewed NRC headquarters and
regional officials and regional and on-site inspectors responsible for
implementation of the ROP and visited the Salem and Hope Creek nuclear
power plants to observe firsthand how the ROP is implemented by the
resident inspectors located at each power plant. To learn more about how
ROP results are communicated to the public, we attended the annual public
meeting held at the Indian Point nuclear power plant. We also reviewed the
information NRC makes available to the public on a Web site devoted to ROP
topics, which includes both detailed plant-specific information and
general program information and guidance. In addition, we interviewed
external stakeholder individuals and groups about their experiences with
the ROP, including the NRC Commission Chairman from 1995 to 1999, who is
largely credited with leading the development of the ROP; officials from
the NRC Office of the Inspector General, the Nuclear Energy Institute
(NEI), the Union of Concerned Scientists (UCS), and Greenpeace; and
nuclear power plant managers at six sites.

To examine the results of the ROP over the past several years, we reviewed
the number and types of inspection findings NRC issued, the performance
indicators reported by the plants, and the level of oversight NRC provided
to the plants. Specifically, we obtained and analyzed NRC data on its
inspection findings for 2001 through 2005, the years since implementation
of the ROP for which the data were available for the full year, and
discussed our analysis with NRC program officials. We obtained inspection
findings data from the NRC Reactor Program System (RPS) database and
assessed the reliability of the RPS data by (1) performing electronic and
manual testing of required data elements; (2) reviewing existing
information about the data and the system that produced them; (3)
reviewing an audit of the RPS database performed by the NRC Office of the
Inspector General in 2005, and documents related to NRC's implementation
of the audit's recommendations; and (4) interviewing agency officials
knowledgeable about these data. We determined that these data were
sufficiently reliable for the purposes of this report. Data elements
included a breakout of inspection findings identified at each nuclear
power plant, including information on their risk significance, or color;
cornerstone; date; cross-cutting aspects; a brief description of the
problem identified; and other related information. We also obtained and
analyzed ROP data provided on NRC's Web site, including a list of
performance indicator data broken out by plant and by quarter and a list
of the plants in each column of the action matrix by quarter since
inception of the ROP in 2000. We assessed the reliability of these data by
interviewing the appropriate agency officials about how these data are
reported on the Web site and compared the data with source information
contained in inspection reports and assessment letters for sample plants.
We determined that these data were sufficiently reliable for the purposes
of this report. In addition, we reviewed every assessment letter issued to
each of the 103 plants since the inception of the ROP to document the
plants that had one or more substantive cross-cutting issue open and the
length of time that the issue was held open. We discussed the results of
our data analysis with NRC headquarters program officials and compared the
results of the ROP with other industry-collected and reported performance
data, including data collected through NRC's industry trends program, to
identify any inconsistencies in trends or industry safety performance
indicators.

To examine the status of NRC's efforts to improve the ROP, we reviewed
specific components of the ROP where weaknesses had been identified and
recent and current staffs' efforts to improve them. We analyzed NRC
documents, including all annual self-assessment reports issued by NRC
since 2001; interviewed officials from NRC headquarters, regional, and
site offices and outside stakeholder groups, including NEI and UCS; and
attended two public meetings covering proposed changes to incorporate
safety culture into its oversight process. In reviewing the annual
self-assessment reports, we analyzed comments submitted by both internal
and external stakeholders that were collected during recent surveys. We
analyzed responses submitted by internal staff, including management,
program staff, and regional and site inspectors, and responses from
external stakeholders, including industry, industry organizations, public
interest groups, state and local agencies, and members of the public. We
examined all proposed and final documents related to NRC's safety culture
changes, including inspection manuals, training documents, and other
guidance documents. We also reviewed public comments submitted to NRC on
its safety culture changes, such as those submitted by NEI, UCS, and
meetings held on the topic by the Advisory Committee on Reactor Safeguards
(ACRS). In addition, we assessed various external reports and evaluations
related to the ROP or specific aspects relevant to NRC's oversight, such
as issues surrounding safety culture, including our prior reports, those
of the NRC Office of the Inspector General and of the ACRS. We met with
NRC program officials responsible for assessing and implementing changes
to the ROP to obtain a clear understanding of the actions they were taking
and the status of their efforts.

Additionally, we selected a nonprobability sample1 of 6 nuclear power
sites (totaling 11 of the 103 operating plants)2 that provided coverage of
each of NRC's four regional offices and varying levels of plant
performance and NRC oversight since 2000. The following nuclear power
sites were included in our review: Cooper (1 plant) located near Nebraska
City, Nebraska; Indian Point 2 (1 plant) located near New York, New York;
Oconee (3 plants) located near Greenville, South Carolina; Perry (1 plant)
located near Painesville, Ohio; Salem and Hope Creek (3 plants) located
near Lower Alloways Creek, New Jersey; and Surry (2 plants) located near
Newport News, Virginia. Our selection criteria was designed to represent
geographic diversity, a variety of safety problems in which inspection
findings or performance indicators of higher risk significance (white,
yellow, and red) were issued, trends reflecting both improving and
declining safety performance, and plants that have been subjected to at
least some level of increased oversight since the ROP was implemented. The
purpose of our review was to understand how performance problems were
identified by NRC's oversight process, what caused them, actions taken by
NRC and the licensee in response to the problems, and how NRC documented
their resolution. We analyzed all publicly available inspection reports
and assessment documents covering years 2000 through 2005 for each site to
examine how NRC applied the ROP to identify and correct safety problems.3
We analyzed each green and greater-than-green inspection finding
documented through the inspection reports and collected data associated
with each finding, including how it was identified, which cornerstone it
was assigned to, whether the finding was associated with a violation of
regulatory requirements, and whether there were any cross-cutting elements
associated with the finding. For the greater-than-green inspection
findings, we collected additional information to determine the length of
time it took the licensee to correct the performance problem, the length
of time and level of effort NRC inspectors took to follow up on the issue,
and the actions it required the licensee to take. Additionally, for each
of the sites, we reviewed NRC reports and documentation showing that all
baseline inspection procedures were completed (for 2004 and 2005), and
that inspectors verified the licensees' reporting of the performance
indicator data (for 2000 through 2005). We also interviewed NRC branch
chiefs and resident inspectors and industry management officials at each
site to learn more about NRC's implementation of the ROP at the site. We
conducted our work from July 2005 through July 2006 in accordance with
generally accepted government auditing standards.

Appendix III

Nuclear Power Plant Performance Data on the Basis of the Results of NRC's
Reactor Oversight Process, 2001 Through 2005

Table 3: Commercial Nuclear Power Plants Licensed to Operate in the United
States

                                        

        Nuclear power plant                City           State   NRC region  
Arkansas Nuclear 1             Russellville           AR      IV           
Arkansas Nuclear 2             Russellville           AR      IV           
Beaver Valley 1                McCandless             PA      I            
Beaver Valley 2                McCandless             PA      I            
Braidwood 1                    Joilet                 IL      III          
Braidwood 2                    Joilet                 IL      III          
Browns Ferry 1                 Decatur                AL      II           
Browns Ferry 2                 Decatur                AL      II           
Browns Ferry 3                 Decatur                AL      II           
Brunswick 1                    Southport              NC      II           
Brunswick 2                    Southport              NC      II           
Byron 1                        Rockford               IL      III          
Byron 2                        Rockford               IL      III          
Callaway                       Fulton                 MO      IV           
Calvert Cliffs 1               Annapolis              MD      I            
Calvert Cliffs 2               Annapolis              MD      I            
Catawba 1                      Rock Hill              SC      II           
Catawba 2                      Rock Hill              SC      II           
Clinton                        Clinton                IL      III          
Columbia Generating Station    Richland               WA      IV           
Comanche Peak 1                Glen Rose              TX      IV           
Comanche Peak 2                Glen Rose              TX      IV           
Cooper                         Nebraska City          NE      IV           
Crystal River 3                Crystal River          FL      II           
D.C. Cook 1                    Benton Harbor          MI      III          
D.C. Cook 2                    Benton Harbor          MI      III          
Davis-Besse                    Toledo                 OH      III          
Diablo Canyon 1                San Luis Obispo        CA      IV           
Diablo Canyon 2                San Luis Obispo        CA      IV           
Dresden 2                      Morris                 IL      III          
Dresden 3                      Morris                 IL      III          
Duane Arnold                   Cedar Rapids           IA      III          
Edwin I. Hatch 1               Baxley                 GA      II           
Edwin I. Hatch 2               Baxley                 GA      II           
Fermi 2                        Toledo                 MI      III          
Fort Calhoun                   Omaha                  NE      IV           
Ginna                          Rochester              NY      I            
Grand Gulf 1                   Vicksburg              MS      IV           
H.B. Robinson 2                Florence               SC      II           
Hope Creek 1                   Lower Alloways Creek   NJ      I            
Indian Point 2                 New York               NY      I            
Indian Point 3                 New York               NY      I            
James A. FitzPatrick           Oswego                 NY      I            
Joseph M. Farley 1             Dothan                 AL      II           
Joseph M. Farley 2             Dothan                 AL      II           
Kewaunee                       Green Bay              WI      III          
La Salle 1                     Ottawa                 IL      II           
La Salle 2                     Ottawa                 IL      II           
Limerick 1                     Philadelphia           PA      I            
Limerick 2                     Philadelphia           PA      I            
McGuire 1                      Charlotte              NC      II           
McGuire 2                      Charlotte              NC      II           
Millstone 2                    New London             CT      I            
Millstone 3                    New London             CT      I            
Monticello                     Minneapolis            MN      III          
Nine Mile Point 1              Oswego                 NY      I            
Nine Mile Point 2              Oswego                 NY      I            
North Anna 1                   Richmond               VA      II           
North Anna 2                   Richmond               VA      II           
Oconee 1                       Greenville             SC      II           
Oconee 2                       Greenville             SC      II           
Oconee 3                       Greenville             SC      II           
Oyster Creek                   Toms River             NJ      I            
Palisades                      South Haven            MI      III          
Palo Verde 1                   Phoenix                AZ      IV           
Palo Verde 2                   Phoenix                AZ      IV           
Palo Verde 3                   Phoenix                AZ      IV           
Peach Bottom 2                 Lancaster              PA      I            
Peach Bottom 3                 Lancaster              PA      I            
Perry 1                        Painesville            OH      III          
Pilgrim 1                      Plymouth               MA      I            
Point Beach 1                  Manitowoc              WI      III          
Point Beach 2                  Manitowoc              WI      III          
Prairie Island 1               Minneapolis            MN      III          
Prairie Island 2               Minneapolis            MN      III          
Quad Cities 1                  Moline                 IL      III          
Quad Cities 2                  Moline                 IL      III          
River Bend 1                   Baton Rouge            LA      IV           
Salem 1                        Lower Alloways Creek   NJ      I            
Salem 2                        Lower Alloways Creek   NJ      I            
San Onofre 2                   San Clemente           CA      IV           
San Onofre 3                   San Clemente           CA      IV           
Seabrook 1                     Portsmouth             NH      I            
Seqouyah 1                     Chattanooga            TN      II           
Seqouyah 2                     Chattanooga            TN      II           
Shearon Harris 1               Raleigh                NC      II           
South Texas Project 1          Bay City               TX      IV           
South Texas Project 2          Bay City               TX      IV           
St. Lucie 1                    Ft. Pierce             FL      II           
St. Lucie 2                    Ft. Pierce             FL      II           
Summer                         Columbia               SC      II           
Surry 1                        Newport News           VA      II           
Surry 2                        Newport News           VA      II           
Susquehanna 1                  Berwick                PA      I            
Susquehanna 2                  Berwick                PA      I            
Three Mile Island 1            Harrisburg             PA      I            
Turkey Point 3                 Miami                  FL      II           
Turkey Point 4                 Miami                  FL      II           
Vermont Yankee                 Battleboro             VT      I            
Vogtle 1                       Augusta                GA      II           
Vogtle 2                       Augusta                GA      II           
Waterford 3                    New Orleans            LA      IV           
Watts Bar 1                    Spring City            TN      II           
Wolf Creek 1                   Burlington             KS      IV           

Source: NRC.

Table 4: Total Number of Green Inspection Findings, 2001 Through 2005

                                        

                             Number of green                
                         inspection findings, by            
                                  year                      
Nuclear power sitea                       2001 2002 2003 2004 2005   Total 
Arkansas Nuclear                            12    9   15   23   23      82 
Beaver Valley                               14   10   10    8    9      51 
Braidwood                                   11   12    5    5    7      40 
Browns Ferry                                 5    5    4   10    8      32 
Brunswick                                    4    4    3    6    9      26 
Byron                                       13    9   16   14   19      71 
Callaway                                    17   11   15   17   17      77 
Calvert Cliffs                              18    4    6   19    6      53 
Catawba                                     10    6    7    5    8      36 
Clinton                                      8   10    5   16    8      47 
Columbia Generating                         11    5   21   17   10      64 
Station                                                             
Comanche Peak                                6   12    3    9   11      41 
Cooper                                      32   23   29   23   34     141 
Crystal River                                5    4    5    5    5      24 
D.C. Cook                                   14   27   24    7   16      88 
Davis-Besse                                  4   16   24   46   12     102 
Diablo Canyon                               11   14   22   25   23      95 
Dresden                                     20   19   22   26   13     100 
Duane Arnold                                 1    7   22   12   19      61 
Farley                                       8    5    5    4   11      33 
Fermi                                        6    6   11   14   18      55 
FitzPatrick                                 12    7    8    5    8      40 
Fort Callhoun                                7   18   11   15   21      72 
Ginna                                        4   11   14    9   10      48 
Grand Gulf                                  13    9    6   10   11      49 
Harris                                       5    5    5    5    6      26 
Hatch                                       11    8    4    2    4      29 
Hope Creek                                  18   24   19   25   16     102 
Indian Point 2                              44   21   20   25   13     123 
Indian Point 3                              13    3   14    6    8      44 
Kewaunee                                    12    9   14   24   20      79 
La Salle                                     9   13    7   11   20      60 
Limerick                                    15   11   13    8    4      51 
McGuire                                      5    9    8   14    9      45 
Millstone                                   20   13    8   12   17      70 
Monticello                                  17    8   12   13    9      59 
Nine Mile Point                             16    9   15   19    8      67 
North Anna                                   4    4    5    9   12      34 
Oconee                                      16   11   13   11   19      70 
Oyster Creek                                11   12    7   15   12      57 
Palisades                                   23   13   13   10    8      67 
Palo Verde                                  16    3    4   46   29      98 
Peach Bottom                                14   13   18    7    9      61 
Perry                                        7   10   18   16   59     110 
Pilgrim                                      8    4    9    6    7      34 
Point Beach                                 15   10   28   24   23     100 
Prarie Island                                4    6    5   10   16      41 
Quad Cities                                  6   21   15   14   11      67 
River Bend                                  19    6   17   18    7      67 
Robinson                                     4    1    4    3    3      15 
Saint Lucie                                 13    7    6   10    9      45 
Salem                                       17   14   26   24   25     106 
San Onofre                                   4    9    9   16   13      51 
Seabrook                                    10    6   10    9    8      43 
Sequoyah                                    11   16    9    7    4      47 
South Texas                                 14    7   15   13   12      61 
Summer                                      13    9   10   10    9      51 
Surry                                        6    5    7    3    7      28 
Susquehanna                                 11   17   11   13    7      59 
Three Mile Island                           19    8    7   12   12      58 
Turkey Point                                 6    2    9   13    9      39 
Vermont Yankee                              11    9    9   14    3      46 
Vogtle                                       4    8    6    8    3      29 
Waterford                                    5   14   15   17   15      66 
Watts Bar                                    4   13   10    8    8      43 
Wolf Creek                                   5    3    7    9    6      30 
Total                                      751  657  774  889  835   3,906 
Site average                                11   10   12   13   13      59 
Range (for each                           1-44 1-27 3-29 2-46 3-59  15-141 
column)                                                             

Source: GAO analysis of NRC data.

aNRC reports these data by nuclear power site as opposed to by individual
plant. Oftentimes, there are 2 or 3 plants located at each site.
Therefore, data for all 103 plants are included here, but at the site
level.

Table 5: Total Number of Greater-Than-Green Inspection Findings Issued,
2001 Through 2005

                                        

                                    Total number of inspection     
                                        findings, by color         
       Nuclear power sitea                                   White Yellow Red 
Arkansas Nuclear                                              1        
Beaver Valley                                                 2        
Braidwood                                                     1        
Browns Ferry                                                           
Brunswick                                                     1        
Byron                                                                  
Callaway                                                      3        
Calvert Cliffs                                                3      1 
Catawba                                                                
Clinton                                                       1        
Columbia Generating Station                                   1      1 
Comanche Peak                                                 2        
Cooper                                                        5        
Crystal River                                                 1        
D.C. Cook                                                     3        
Davis-Besse                                                   4      1   1 
Diablo Canyon                                                          
Dresden                                                       2        
Duane Arnold                                                           
Farley                                                                 
Fermi                                                         1        
FitzPatrick                                                            
Fort Callhoun                                                 2        
Ginna                                                         1        
Grand Gulf                                                             
Harris                                                        2        
Hatch                                                         1        
Hope Creek                                                    2        
Indian Point 2                                                2      1   1 
Indian Point 3                                                         
Kewaunee                                                      4      1 
La Salle                                                      1        
Limerick                                                      2        
McGuire                                                                
Millstone                                                              
Monticello                                                             
Nine Mile Point                                               1        
North Anna                                                             
Oconee                                                        7        
Oyster Creek                                                  3        
Palisades                                                     2        
Palo Verde                                                           1 
Peach Bottom                                                  3        
Perry                                                         5        
Pilgrim                                                                
Point Beach                                                   3      1   3 
Prarie Island                                                 1        
Quad Cities                                                            
River Bend                                                    2        
Robinson                                                               
Saint Lucie                                                            
Salem                                                         1        
San Onofre                                                             
Seabrook                                                      1        
Sequoyah                                                      1        
South Texas                                                            
Summer                                                                 
Surry                                                         2        
Susquehanna                                                   1        
Three Mile Island                                             2        
Turkey Point                                                           
Vermont Yankee                                                1        
Vogtle                                                                 
Waterford                                                     1        
Watts Bar                                                     1        
Wolf Creek                                                             
Total                                                        86      7   5 

Source: GAO analysis of NRC data.

aNRC reports these data by nuclear power site as opposed to by individual
plant. Oftentimes, there are 2 or 3 plants located at each site.
Therefore, data for all 103 plants are included here, but at the site
level.

Table 6: Type of Substantive Cross-cutting Issue Open At Least Some
Portion of the Year, 2001 Through 2005

                                        

                       Type of substantive  
                       cross-cutting issue  
                          open, by year     
     Nuclear power            2001           2002   2003     2004      2005   
         plant                                                       
Arkansas Nuclear                                PIR     PIR       PIR      
1                                                                 
Arkansas Nuclear                                PIR     PIR       PIR      
2                                                                 
Beaver Valley 1                                                   
Beaver Valley 2                                                   
Braidwood 1                                                       
Braidwood 2                                                       
Browns Ferry 2                                                    
Browns Ferry 3                                                    
Brunswick 1                                                       
Brunswick 2                                                       
Byron 1                                                           HP       
Byron 2                                                           HP       
Callaway                                 PIR    PIR     HP        HP       
Calvert Cliffs 1   PIR                   PIR                      
Calvert Cliffs 2   PIR                   PIR                      
Catawba 1                                                         
Catawba 2                                                         
Clinton                                                           
Columbia           HP                    HP     PIR, HP PIR, HP   PIR, HP  
Generating                                                        
Station                                                           
Comanche Peak 1                                                   
Comanche Peak 2                                                   
Cooper             PIR,HP                PIR,HP PIR,HP  PIR,HP    PIR,HP   
Crystal River 3                                                   
D.C. Cook 1                              PIR    PIR               
D.C. Cook 2                              PIR    PIR               
Davis-Besse                                                       
Diablo Canyon 1                                 PIR,HP  PIR,HP    
Diablo Canyon 2                                 PIR,HP  PIR,HP    
Dresden 2                                               HP        
Dresden 3                                               HP        
Duane Arnold                                                      HP       
Farley 1                                                          
Farley 2                                                          
Fermi 2                                                 HP        HP       
FitzPatrick        PIR                                            
Fort Callhoun                                                     
Ginna                                                             
Grand Gulf 1                                                      
Harris 1                                                          
Hatch 1                                                           
Hatch 2                                                           
Hope Creek 1                             PIR    PIR     PIR, SCWE PIR,SCWE 
Indian Point 2     PIR,HP                PIR,HP PIR,HP  PIR       PIR      
Indian Point 3                                                    
Kewaunee                                                PIR       PIR      
La Salle 1         HP                    HP             HP        HP       
La Salle 2         HP                    HP             HP        HP       
Limerick 1                                                        
Limerick 2                                                        
McGuire 1                                                         
McGuire 2                                                         
Millstone 2        PIR,HP                       PIR               
Millstone 3                                                       
Monticello                                                        
Nine Mile Point 1                               PIR               
Nine Mile Point 2                               PIR               
North Anna 1                                                      
North Anna 2                                                      
Oconee 1                                                          
Oconee 2                                                          
Oconee 3                                                          
Oyster Creek       PIR                   HP     HP      PIR       PIR      
Palisades          PIR, HP               PIR    PIR               
Palo Verde 1                                            PIR, HP   PIR, HP  
Palo Verde 2                                            PIR, HP   PIR, HP  
Palo Verde 3                                            PIR, HP   PIR, HP  
Peach Bottom 2                                  PIR     PIR       
Peach Bottom 3                                  PIR     PIR       
Perry 1                                                 PIR,HP    PIR,HP   
Pilgrim 1                                                         
Point Beach 1                            PIR    PIR, HP PIR, HP   PIR, HP  
Point Beach 2                            PIR    PIR, HP PIR, HP   PIR, HP  
Prairie Island 1                                                  
Prairie Island 2                                                  
Quad Cities 1                            HP     HP                
Quad Cities 2                            HP     HP                
River Bend 1                                                      
Robinson 2                                                        
Saint Lucie 1                                                     
Saint Lucie 2                                                     
Salem 1                                  PIR    PIR     PIR,SCWE  PIR,SCWE 
Salem 2                                  PIR    PIR     PIR,SCWE  PIR,SCWE 
San Onofre 2                                                      
San Onofre 3                                                      
Seabrook 1         PIR                   PIR                      
Sequoyah 1                                                        
Sequoyah 2                                                        
South Texas 1                                                     
South Texas 2                                                     
Summer                                                            
Surry 1                                                           
Surry 2                                                           
Susquehanna 1                            HP     PIR, HP PIR       
Susquehanna 2                            HP     PIR, HP PIR       
Three Mile Island  HP                    HP     PIR     PIR       
1                                                                 
Turkey Point 3                                  PIR               
Turkey Point 4                                  PIR               
Vermont Yankee                                                    
Vogtle 1                                                          
Vogtle 2                                                          
Waterford 3                              PIR    PIR               
Watts Bar 1                                                       HP       
Wolf Creek 1                                                      

Legend:

HP       =  human performance PIR =  problem identification and resolution
SCWE =  safety-conscious work environment

Source: GAO analysis of NRC data.

Table 7: Total Number of Greater-Than-Green Performance Indicators, 2001
Through 2005

                                        

                      Number of greater-than-green           
                       performance indicators, by            
                                 yeara                       
     Nuclear power                            2001 2002 2003 2004 2005  Total 
         plant                                                          
Arkansas Nuclear                                                     
1                                                                    
Arkansas Nuclear                                                     
2                                                                    
Beaver Valley 1                                                      
Beaver Valley 2                                                      
Braidwood 1                                   1    2    3    1           7 
Braidwood 2                                                          
Browns Ferry 2                                                       
Browns Ferry 3                                                       
Brunswick 1                                                          
Brunswick 2                                                       1      1 
Byron 1                                                              
Byron 2                                                              
Callaway                                                     2           2 
Calvert Cliffs 1                              4    3                     7 
Calvert Cliffs 2                                                     
Catawba 1                                                            
Catawba 2                                                            
Clinton                                                              
Columbia                                                          5      5 
Generating                                                           
Station                                                              
Comanche Peak 1                                                      
Comanche Peak 2                                                      
Cooper                                                  1                1 
Crystal River 3                               1                          1 
D.C. Cook 1                                   1                          1 
D.C. Cook 2                                             4    5    2     11 
Davis-Besse                                                  1    1      2 
Diablo Canyon 1                                                   1      1 
Diablo Canyon 2                                                   1      1 
Dresden 2                                                    1           1 
Dresden 3                                          1    4    2           7 
Duane Arnold                                                         
Farley 1                                                             
Farley 2                                                             
Fermi 2                                       4         1    4    1     10 
FitzPatrick                                   2                          2 
Fort Callhoun                                           1    2    2      5 
Ginna                                                                
Grand Gulf 1                                                         
Harris 1                                           1    3    1           5 
Hatch 1                                                              
Hatch 2                                                              
Hope Creek 1                                                         
Indian Point 2                                                       
Indian Point 3                                          3                3 
Kewaunee                                                             
La Salle 1                                                           
La Salle 2                                    2    1                     3 
Limerick 1                                                           
Limerick 2                                                           
McGuire 1                                                            
McGuire 2                                                            
Millstone 2                                   2              3           5 
Millstone 3                                                          
Monticello                                                           
Nine Mile Point 1                             4    2                     6 
Nine Mile Point 2                             1    1    1                3 
North Anna 1                                                         
North Anna 2                                  1                          1 
Oconee 1                                           2                     2 
Oconee 2                                                             
Oconee 3                                                             
Oyster Creek                                                         
Palisades                                                            
Palo Verde 1                                                         
Palo Verde 2                                                         
Palo Verde 3                                                         
Peach Bottom 2                                          1    1    3      5 
Peach Bottom 3                                                       
Perry 1                                                 3    1           4 
Pilgrim 1                                                            
Point Beach 1                                 4    1                     5 
Point Beach 2                                 2                          2 
Prarie Island 1                                                      
Prarie Island 2                                                      
Quad Cities 1                                                        
Quad Cities 2                                                        
River Bend 1                                                      2      2 
Robinson 2                                                   1           1 
Saint Lucie 1                                                        
Saint Lucie 2                                           1    1           2 
Salem 1                                            2                     2 
Salem 2                                                              
San Onofre 2                                            1    3    1      5 
San Onofre 3                                                         
Seabrook 1                                              1                1 
Sequoyah 1                                                           
Sequoyah 2                                         1    2                3 
South Texas 1                                                        
South Texas 2                                      1    1                2 
Summer                                        3                          3 
Surry 1                                            3    6    3          12 
Surry 2                                       2    4    4    1          11 
Susquehanna 1                                                        
Susquehanna 2                                                        
Three Mile Island                                                    
1                                                                    
Turkey Point 3                                                    1      1 
Turkey Point 4                                                    2      2 
Vermont Yankee                                                       
Vogtle 1                                                             
Vogtle 2                                                             
Waterford 3                                                          
Watts Bar 1                                                          
Wolf Creek 1                                                         
Total                                        34   25   41   33   23    156 

Source: GAO analysis of NRC data.

Note: Plant licensees report their performance indicator data on a
quarterly basis for 15 different indicators (excluding 3 physical security
indicators). Yearly totals include a summary for all 15 indicators and the
four quarters. Thus, if the same indicator was white for two quarters
during the year, it would count twice in the yearly total.

aAll of the greater-than-green indicators were white during this period,
no yellow or red indicators were reported.

Table 8: Highest NRC Oversight Level Applied during at Least Some Portion
of the Year, 2001 Through 2005

                                        

                           Highest level of oversight     
                          applied during at least some    
                              portion of the yeara        
Nuclear power plant                2001                2002 2003 2004 2005 
Arkansas Nuclear 1                                               L    
Arkansas Nuclear 2                                                    
Beaver Valley 1                                        L    L    L    
Beaver Valley 2                                        L    L    L    
Braidwood 1          L                                 M    L    L    
Braidwood 2                                                           
Browns Ferry 2                                                        
Browns Ferry 3                                                        
Brunswick 1                                                           
Brunswick 2                                                      L    L    
Byron 1                                                               
Byron 2                                                               
Callaway             M                                 L    L    L    
Calvert Cliffs 1     M                                 M    L         
Calvert Cliffs 2                                       L    L    L    
Catawba 1                                                             
Catawba 2                                                             
Clinton              L                                 L              
Columbia Generating  M                                 M              L    
Station                                                               
Comanche Peak 1                                        L         L    
Comanche Peak 2                                        L              
Cooper               M                                 H    H    H    
Crystal River 3      L                                                L    
D.C. Cook 1                                            L    L    L    
D.C. Cook 2                                            M    M    M    L    
Davis-Besse                                            b    b    b    b    
Diablo Canyon 1                                                       L    
Diablo Canyon 2                                                       L    
Dresden 2                                                        L    
Dresden 3                                              L    L    L    
Duane Arnold                                                          
Farley 1             L                                                
Farley 2             L                                                
Fermi 2              L                                 L    L    L    L    
FitzPatrick          L                                                
Fort Callhoun                                          L    L    L    L    
Ginna                                                  L    L         
Grand Gulf 1                                                          
Harris 1             L                                 M    L    L    
Hatch 1                                                               L    
Hatch 2                                                               L    
Hope Creek 1                                                     L    L    
Indian Point 2       H                                 H    M    L    L    
Indian Point 3                                              L         
Kewaunee             M                                 L    L         M    
La Salle 1                                                            
La Salle 2           L                                 L              
Limerick 1           L                                 L              
Limerick 2           L                                 L              
McGuire 1                                                             
McGuire 2                                                             
Millstone 2          M                                           L    
Millstone 3                                                           
Monticello                                                            
Nine Mile Point 1    L                                 L    L         
Nine Mile Point 2    L                                 L    L         
North Anna 1                                                          
North Anna 2         L                                                
Oconee 1             M                                 L    L    M    L    
Oconee 2             L                                      L    M    L    
Oconee 3             L                                 L    L    M    L    
Oyster Creek         L                                 L         L    L    
Palisades            L                                 L    L    L    
Palo Verde 1                                                          M    
Palo Verde 2                                                          M    
Palo Verde 3                                                          M    
Peach Bottom 2       L                                 L    L    L    L    
Peach Bottom 3       L                                 L    L         
Perry 1                                                L    M    H    H    
Pilgrim 1                                                             
Point Beach 1        L                                 L    H    H    H    
Point Beach 2        L                                 L    H    H    H    
Prarie Island 1      L                                 L              
Prarie Island 2      L                                 L              
Quad Cities 1        L                                                
Quad Cities 2        L                                                
River Bend 1                                           L              L    
Robinson 2                                                       L    
Saint Lucie 1                                                         
Saint Lucie 2                                               L    L    
Salem 1                                                L    L    L    
Salem 2                                                               
San Onofre 2                                                L    L    L    
San Onofre 3                                                          
Seabrook 1           L                                      L         
Sequoyah 1                                                       L    L    
Sequoyah 2                                             L    L         
South Texas 1                                                         
South Texas 2                                          L    L         
Summer               L                                                
Surry 1              L                                 L    L    M    
Surry 2              L                                 L    L    L    
Susquehanna 1        L                                 L              
Susquehanna 2        L                                 L              
Three Mile Island 1  L                                                L    
Turkey Point 3                                                        L    
Turkey Point 4                                                        L    
Vermont Yankee                                         M         L    L    
Vogtle 1                                                              
Vogtle 2                                                              
Waterford 3                                                      L    
Watts Bar 1                                                           L    
Wolf Creek 1                                                          

Legend:

L  = 1st or lowest level of increased oversight beyond the baseline
M = 2nd level of increased oversight beyond the baseline H  = 3rd and
highest level of oversight that still allows continued plant operations

Source: GAO analysis of NRC data.

aAll plants receive a baseline level of oversight, regardless of their
safety performance. Plants are also placed into performance categories on
NRC's action matrix on a quarterly basis, which corresponds to the level
of oversight NRC will provide based on the plant's safety performance. The
level of oversight reported here corresponds to the highest oversight
level the plant received during the year, even if it was only for a
portion of the year. Thus, if a plant was placed into a different category
each quarter, the highest category in which it was placed is reported
here.

bDavis-Besse was under a separate oversight category not considered part
of the ROP due to the reactor vessel head incident that occurred in 2002.

Appendix IV

Comments from the Nuclear Regulatory Commission

Appendix V

GAO Contact and Staff Acknowledgments

GAO Contact

Jim Wells, (202) 512-3841 or [email protected]

Staff Acknowledgments

In addition to the individual named above, Raymond H. Smith, Jr.
(Assistant Director), Alyssa M. Hundrup, and Dave Stikkers made key
contributions to this report. Also contributing to this report were Cindy
Gilbert, Carol Kolarik, Alison O'Neill, Ilene Pollack, Keith A. Rhodes,
and Barbara Timmerman.

(360585)

www.gao.gov/cgi-bin/getrpt? GAO-06-1029 .

To view the full product, including the scope
and methodology, click on the link above.

For more information, contact Jim Wells at (202) 512-3841 or
[email protected].

Highlights of GAO-06-1029 , a report to congressional requesters

September 2006

NUCLEAR REGULATORY COMMISSION

Oversight of Nuclear Power Plant Safety Has Improved, but Refinements Are
Needed

The Nuclear Regulatory Commission (NRC) is responsible for overseeing the
nation's 103 commercial nuclear power plants to ensure they are operated
safely. The safety of these plants has always been important, since an
accident could release harmful radioactive material. NRC's oversight has
become even more critical as the potential resurgence of nuclear power is
considered. NRC implemented a new Reactor Oversight Process (ROP) in 2000
to address weaknesses in its oversight of nuclear plant safety.

In this report, GAO reviewed (1) how NRC oversees nuclear power plants,
(2) the results of the ROP over the past several years, and (3) the status
of NRC's efforts to improve the ROP. To complete this work, GAO analyzed
programwide information, inspection results covering 5 years of ROP
operations, and detailed findings from a nonprobability sample of 11
plants.

What GAO Recommends

GAO recommends that NRC aggressively monitor; evaluate; and, if needed,
implement additional measures to increase the effectiveness of its safety
culture changes and make publicly available more information on nuclear
power plants' safety culture. In commenting on a draft of this report, NRC
generally agreed with GAO's findings, conclusions, and recommendations.

NRC uses various tools and takes a risk-informed and graded approach to
ensure the safety of nuclear power plants. These tools consist of physical
inspections of plants and quantitative measures or indicators of plant
performance. They are risk-informed in that they focus on the aspects of
operations considered most important to plant safety. On the basis of the
results of this information, NRC takes a graded approach to its oversight,
increasing the level of regulatory attention to plants where safety is
declining. NRC assesses overall plant performance and communicates the
results to the public, including providing detailed results of its
oversight process through a Web site devoted to the ROP.

Since 2001, the ROP has resulted in more than 4,000 inspection findings
concerning plants' failure to fully comply with safe operating procedures,
and NRC has subjected more than 75 percent (79) of the 103 plants to
increased oversight for varying periods. Almost all of the inspection
findings were for actions NRC considered important to correct but of low
significance to safe plant operations. While the majority of plants
received some level of increased oversight, only 5 plants were subjected
to NRC's highest level of oversight. Plants in this category were
generally subjected to this higher oversight for long periods due to the
more systemic nature of their performance problems.

NRC has improved its oversight process in various areas, but it has been
slow to act on needed improvements, particularly in improving the agency's
ability to identify and address early indications of declining safety
performance. NRC is improving its oversight process on the basis of
feedback from stakeholders, including better focusing inspections on areas
most important to safety. NRC also is addressing what GAO believes has
been a significant shortcoming by modifying the ROP to improve its ability
to address plants' safety culture-that is, the organizational
characteristics that ensure issues affecting nuclear plant safety receive
the attention their significance warrants. GAO and others, including some
stakeholders, believe these changes could enable NRC to better identify
safety culture issues and thus provide earlier indications of declining
plant safety performance. However, some in the industry have opposed the
changes because they believe the changes could introduce undue
subjectivity to NRC's oversight, given the difficulty in measuring these
often intangible and complex concepts. NRC has been reluctant to
incorporate safety culture into the ROP because it considered this type of
activity as a management function, and NRC did not believe that it should
be directly involved in managing licensees' plants. NRC program officials
view the current changes as the beginning of an incremental approach and
acknowledge that they will need to assess their effectiveness at
identifying declining safety performance at plants before significant
safety events occur.

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 ( 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
www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud, Waste, and Abuse in Federal Programs

Contact:

Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: [email protected]
Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional Relations

Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548

Public Affairs

Paul Anderson, Managing Director, [email protected] (202) 512-4800 U.S.
Government Accountability Office, 441 G Street NW, Room 7149 Washington,
D.C. 20548
*** End of document. ***