ASSESSMENT OF THE INCENTIVES
CREATED BY PUBLIC DISCLOSURE
OF OFF-SITE CONSEQUENCE ANALYSIS INFORMATION
FOR REDUCTION IN THE RISK OF ACCIDENTAL RELEASES
i © s
v*y
cepNH
April 18, 2000
United States Environmental Protection Agency
Ariel Rios Building
1200 Pennsylvania Avenue, NW
Washington, DC 20460
-------
This page intentionally left blank.
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY 1
CHAPTER 1- INTRODUCTION 5
CHAPTER 2 - ARE CHEMICAL ACCIDENTS A SIGNIFICANT PROBLEM? 11
CHAPTER 3 - DOES PUBLIC INFORMATION LEAD TO RISK REDUCTION? 16
CHAPTER 4 - DOES THE TYPE OF INFORMATION AND TYPE OF ACCESS MAKE A
DIFFERENCE? 25
CHAPTER 5 - ARE THERE OTHER SOURCES OF THE SAME DATA? 29
CHAPTER 6 - HOW MUCH INFORMATION IS NECESSARY TO SPARK RISK
REDUCTION EFFORTS? 35
CHAPTER 7 - WHAT IS THE PUBLIC'S ACCESS TO OCA INFORMATION
UNDER CSISSFRRA? 42
CHAPTER 8 - FINDINGS 47
REFERENCES 49
APPENDIX A - BACKGROUND OF EPA'S RISK MANAGEMENT PROGRAM AND
OCA INFORMATION A-l
APPENDIX B - DISCUSSION OF THE APPLICABLE PROVISIONS OF THE
CHEMICAL SAFETY INFORMATION, SITE SECURITY, AND FUELS
REGULATORY RELIEF ACT (CSISSFRRA) B-l
APPENDIX C - USES OF RIGHT-TO-KNOW INFORMATION C-l
APPENDIX D - HOW DOES NEGATIVE PRESS AFFECT
FACILITIES' TOXIC RELEASE EMISSIONS? D-l
APPENDIX E - ANALYSIS OF ACCIDENT DATA E-l
APPENDIX F - INSTITUTIONAL USES OF OCA DATA
CREATING INCENTIVES FOR RISK REDUCTION F-l
-------
LIST OF TABLES
Table 1 - Five-Year Accident History Data from RMP Submissions 12
Table 2 - Dollar Values of Impacts of Releases in the RMP Five-Year Accident History 14
Table 3 - Company-Wide Toxic Release Emissions Reductions 1990-1996 19
Table 4 - Counties with the Most Facilities Submitting an RMP 37
Table 5 - Distribution of RMPs and Counties 37
Table 6 - Industry Sectors Submitting the Most RMPs 38
Table 7 - Distribution of RMPs Across NAICS Codes 39
Table 8 - Chemicals Most Often Reported in RMP WCS and 5-Year Accident History 40
Table 9 - RMP 5-Year Accident History by Industry Sector 40
Table A-l - Data Reported in OCA Sections of an RMP A-11
Table D-l - Reductions of TRI Emissions for Selected Companies D-4
Table D-2 - TRI Reductions for Selected Facilities in Louisiana D-8
Table E-l - RMP Accidents Reported by Industry Sector E-2
Table E-2 - All Hazardous Substance Releases E-4
Table E-3 - Releases of Current TRI Chemicals from Manufacturers E-4
Table E-4 - Releases of Hazardous Substances Above the Current Reportable Quantity .... E-5
Table E-5 - Total Hazardous Substance Releases for Four States E-5
Table E-6 - Oil Releases E-6
Table E-7 - Transportation Releases E-6
Table F-l - Summary of the Natures of Interests in OCA Information F-19
LIST OF EXHIBITS
Exhibit 1 - Episodic Hazardous Substance Releases - Four States 22
Exhibit 2 - Oil and Transportation Releases - Four States 23
Exhibit 3 - Impact of Access and Type of Data 27
Exhibit 4 - Typical Plume Map 32
-------
EXECUTIVE SUMMARY
Under the Chemical Safety Information, Site Security and Fuels Regulatory Relief Act, the
President delegated to the Administrator of the U.S. Environmental Protection Agency (EPA) the
task of assessing the incentives for reduction in accidental chemical releases created by public
disclosure of off-site consequence analysis information. This document reports the results of
EPA's assessment.
In the wake of the chemical tragedy in Bhopal, India, and a series of large chemical
accidents in the U.S. in the late 1980s, Congress added new provisions to the Clean Air Act for
the prevention of accidental chemical releases. In particular, Congress directed EPA to require
facilities that pose the greatest risk of harm to the public and the environment as a result of an
accidental chemical release prepare and submit risk management plans (RMPs). An RMP must
describe the facility's chemical accident prevention program, emergency response program, and
off-site consequence analysis (OCA). The OCA must evaluate the potential for hypothetical
worst-case and alternative accidental release scenarios to harm the public and environment around
the facility. Congress mandated that RMPs be available to state and local governments and the
public.
EPA promulgated RMP requirements in June 1996; the first RMPs were due three years
later. To reduce paperwork burden and to take advantage of today's technology, EPA designed
software tools and forms so that all RMPs could be submitted electronically to EPA and stored in
a central information system. All levels of government would have immediate access to the
system and the most recent RMP submissions. The vast majority of RMPs have been submitted
electronically to EPA.
EPA originally planned to place the RMP information system on the Internet for easy
access by the public, as well as by governments, based on the recommendation of many members
of a Subcommittee created under the Federal Advisory Committee Act. However, concerns were
raised that Internet access to a large, searchable database of OCA results could be used as a
targeting tool by terrorists and other criminals. Although EPA subsequently decided not to place
the OCA sections of RMPs on the Internet, new concerns were raised that recent amendments to
the Freedom of Information Act (FOIA) would compel EPA to release this information in
electronic format. Congress responded by passing the Chemical Safety Information, Site Security
and Fuels Regulatory Relief Act (CSISSFRRA), which the President signed on August 5, 1999.
CSISSFRRA temporarily exempts OCA information from public disclosure under FOIA.
It requires the President to "assess the increased risk of terrorist and other criminal activity
associated with the posting of [OCA] information on the Internet, and the incentives created by
public disclosure of OCA information to reduce the risk of accidental chemical releases." Based
on these assessments, the President is to issue regulations "governing the distribution of [OCA]
information in a manner that, in the opinion of the President, minimizes the likelihood of
accidental releases and [any increased risk of terrorist activity associated with Internet posting of
1
-------
OCA information] and the likelihood of harm to public health and welfare."
FINDINGS OF INCENTIVES ASSESSMENT
* Public disclosure of RMPs including OCA information would likely lead to
significant reduction in the number and severity of accidental chemical releases. In
addition, widespread access to OCA information by all stakeholders would serve the
function Congress originally intended in the Clean Air Act Amendments — to
inform members of the public and allow them to participate in decisions that affect
their lives and communities. The public is not likely to generate such information
on its own, and thus the greater the public access to OCA information, the more
likely potential public safety benefits would be realized.
* Multiple segments of the public, particularly citizens, citizens' groups, and the
media, are likely to become more interested in chemical safety and chemical release risk
reduction, to the extent they become aware of the potentially large consequences
associated with worst-case scenarios and, to a lesser extent, alternative release scenarios.
The interest and concern about potential consequences will likely trigger comparisons and
detailed analyses of not only OCA information but safety and environmental performance
of facilities as well. Widespread awareness of the comparisons and analyses would likely
lead industry to make changes and would stimulate dialogue among facilities, the public,
and local officials to reduce chemical accident risks.
Chemical accidents continue to impose considerable costs in terms of human lives and
health, property damage, and public welfare. Facilities covered by the RMP rule reported
that from mid-1994 to mid-1999 there were about 1,900 serious accidents that caused 33
deaths, 8,300 injuries, and the evacuation or sheltering of 221,000 people. These
accidents cost the affected facilities more than $1 billion in direct damages and two to four
times that in business interruption losses. Almost 80% of these accidents occurred at
facilities already subject to the OSHA process safety management standard, which is
designed to reduce accidents. These accidents also represent less than 10 percent of all
unintended releases of hazardous substances reported to the government during this
period. Additional efforts are needed in order to reduce the number and severity of
chemical accidents.
* Given the opportunity, the public uses hazard information to take action that leads
to risk reduction. Various segments of the public have strong incentives to use OCA
information in ways that reduce risk. For example, there is a broad consensus that
national publication of the Toxics Release Inventory (TRJ) data by the government,
followed by analysis by citizens' groups and the news media, led to action by industry to
reduce emissions. Nationally, reported TRJ emissions have fallen 43 percent since 1988, a
time in which industrial production has risen 28 percent. Although other factors likely
contributed to the decline in emissions, negative press coverage directed at certain
2
-------
facilities appears to have led these facilities to achieve reductions in their TRI emissions.
It is not possible to quantify the exact level of risk reduction that would be gained from
public dissemination of OCA information, but the effect would likely be significant.
• Ease of access to information is important to public use and risk reduction. Data
available in paper form on request from state or local agencies are rarely sought. For
example, data on the location and identity of hazardous chemicals are requested about
3,500 times a year from Local Emergency Planning Committees (LEPCs). (There are
about 3,200 LEPCs in the country and about 560,000 facilities subject to requirements to
report information on hazardous chemicals to LEPCs.) Meanwhile, environmental data on
Environmental Defense's "Scorecard" website are at least 250 times more likely to be
reviewed by the public than information from LEPCs. Likewise, early indications are that
the meetings which facilities were required to conduct by CSISSFRRA to explain OCA
information to the public have drawn very few attendees, even when citizens received
individual invitations. In contrast, when industry has gone out to places the public already
frequents (for example, a shopping mall) and provided consequence information directly to
citizens, outreach and communication about chemical accident risks have been more
successful.
Information that puts hazards into context, as OCA data do, is far more likely to be
used by the public than "raw" data. The importance of such "interpreted" information
(already analyzed in order to be understandable) is demonstrated by the increased use of
TRI data when they were made available as part of Scorecard on the Internet. Although
TRI data are available electronically through EPA's Envirofacts and the RTK-Net (Right-
To-Know Network) websites, Scorecard ranks each facility on various indicators by
county, state, and nation, and explains the health effects of chemicals emitted by that
facility. The raw TRI data on RTK-Net were drawing 240,000 searches a year; Scorecard
draws over a half million page views per month.
• Although OCA data could be derived from other available data, the public is
unlikely to do so. Derivation of OCA data requires some technical knowledge and time.
While motivated and skilled individuals and organizations can use widely available existing
data, guidance, and models to estimate off-site consequences with relative ease, evidence
suggests that the general public is unlikely to be able and willing to do so.
• A complete RMP containing OCA information is necessary to understand the extent
of the hazard posed by a particular facility in comparison to other facilities in an area,
within an industrial sector, or handling the same chemicals. While the OCA data address
the hazard, the RMP information addresses the steps to control those hazards.
Understanding the extent of a hazard and how it is controlled leads to understanding the
risk posed by that facility.
• The penalties for disclosure contained in CSISSFRRA are having a chilling effect,
3
-------
even though the statute provides for access to OCA information for state and local
officials, including emergency planners and responders, and allows those officials to
communicate OCA data to the public. Many of these officials are not willing to obtain or
to communicate the data and thereby to risk accidental or inadvertent disclosure of OCA
information, even though CSISSFRRA penalizes only its willful disclosure. More
fundamentally, making the provision of OCA data to the public discretionary leaves in the
hands of government the decision about whether and to what extent to convey the data.
CSISSFRRA also allows facilities to release their OCA information to the public, but that,
too, is at their discretion. CSISSFRRA's requirement for facilities to conduct a public
meeting or post a public notice summarizing OCA information provided only a one-time
opportunity to learn about local hazards.
• Actual chemical releases are different from the releases evaluated for OCA purposes.
No one can control all of the conditions (for example, weather) used to develop an off-site
consequence analysis; actual conditions at a facility can vaiy widely from those used in the
analysis. The accident prevention rule requires facilities to conduct OCAs in a specified,
systematic manner so that the public and others can understand the relative hazards and
risks posed by facilities as a result of the type and amount of chemical handled and the
mitigation measures used.
This assessment finds that convenient public access to OCA information has the power to
reduce real impact associated with chemical accidents. America needs to be further educated
about chemical risks. Dissemination of OCA information could make an important contribution
to a public dialogue about risk reduction and protection of lives. This public dialogue among
community members, emergency planners and responders, and facilities at the local level is key to
risk reduction.
4
-------
CHAPTER 1
INTRODUCTION
"/ know of no safe depository of the ultimate powers of society but the people
themselves; and if we think them not enlightened enough to exercise their control
with a wholesome discretion, the remedy is not to take it from them, but to inform
their discretion" (Thomas Jefferson, letter to William Charles Jarvis, September
28, 1820).
The federal government's efforts to prevent and mitigate chemical accidents have come
largely in the wake of the 1984 accidental chemical release in Bhopal, India, that killed more than
3,000 people and injured more than 100,000 (1). This incident demonstrated to the world the
magnitude of the potential consequences of a single chemical accident.
But this was not an isolated event. Less than one month prior to the Bhopal accident, an
accidental release of liquefied petroleum gas (i.e., propane) from a storage terminal in Mexico
City resulted in a large fire and series of explosions, killing 500 people and destroying a residential
area (2). Other catastrophic chemical accidents have occurred in countries throughout the world,
including the United States. In 1985, an accident at a Union Carbide plant in Institute, West
Virginia, led to a release of a noxious mixture of methylene chloride and aldicarb oxime, resulting
in the hospitalization of 134 people living in surrounding areas.
As a result, Congress passed the Emergency Planning and Community Rjght-to-Know Act
(EPCRA) in 1986 as a part of the Superfund Amendments and Reauthorization Act. EPCRA
calls on states to create State Emergency Response Commissions (SERCs) and communities to
form Local Emergency Planning Committees (LEPCs) to prepare local emergency response plans
for chemical accidents. EPCRA also requires facilities to provide LEPCs with information
necessary for emergency planning, and to submit to SERCs, LEPCs and local fire departments
annual inventory reports and information about hazardous chemicals. The statute also established
the Toxics Release Inventory (TRI), which requires certain facilities to annually report the
quantities of their emissions of toxic chemicals. These data are to be available to the public and
EPA is to maintain a national database containing these toxic chemical release reports.
However, EPCRA contains no provisions for the prevention of chemical accidents and,
because major accidental releases continued to occur, Congress included two provisions in the
Clean Air Act (CAA) Amendments of 1990 to institute federal regulatory programs to prevent
chemical accidents that harm workers, the public and the environment. Section 304 of the
Amendments calls for chemical accident prevention and emergency response regulations to
protect workers on-site, while section 112(r) of the amended CAA calls for regulations to prevent
and respond to chemical accidents that could affect the public and environment off-site (3).
5
-------
In Section 112(r), Congress established a general duty on facilities handling extremely
hazardous chemicals to do so safely (section 112(r)( 1)), and required EPA to establish regulations
to ensure that facilities that pose the greatest risk develop and implement chemical accident
prevention and detection programs (section 112(r)(7)). Congress further directed that the
chemical accident prevention regulations require that facilities prepare and submit risk
management plans (RMPs); these plans must include a hazard assessment that estimates the
potential consequences of hypothetical worst-case releases, an accident history, a program for
preventing accidental releases, and an emergency response program (section 112(r)(7)(B)(ii)).
Finally, Congress required that these RMPs be submitted to the federal Chemical Safety and
Hazard Investigation Board, state and local emergency response officials, and be made available
to the public (section 112(r)(7)(B)(iii)).
EPA issued a rule in 1994 that lists the
most potentially acutely hazardous toxic and
flammable substances along with a threshold
quantity for each. In 1996 the Agency issued
a rule requiring every facility handling more
than the threshold quantity of a listed
substance to develop and implement a risk
management program based on an assessment
of the hazards at that facility (the "RMP
rule"). As required by section 112(r), EPA
specified in the rule that the hazard assessment
include an off-site consequence analysis
(OCA) of the potential consequences of
worst-case and alternative scenario chemical
releases and that the results of the OCA be
reported in the facility's RMP.
The OCA provides a rough estimate of
the potential consequences to a surrounding
community of one or more hypothetical
accidental releases, without evaluating the
likelihood or probability of such an accident
occurring. Potential consequences are
expressed in terms of potentially exposed
population, as well as the types of buildings,
parks, and other public and environmental
areas that could be seriously affected by a
release.
Chemical Emergency Preparedness and
Prevention - Legislative and Regulatory History
1986 - Emergency Preparedness and Community
Right-to-Knovv Act (EPCRA); PL99-499
1987 - Extremely Hazardous Substances List and
emergency planning and reporting
requirements (40 CFR 355 and 370).
1988 - Toxic Release Inventory (TRI) reporting
requirements (40 CFR 372).
1990 - Clean Air Act (CAA) Amendments,
containing Sections 112(r) and 304; PL 101-
549
1992 - Occupational Safety and Health
Administration (OSHA) Process Safety
Management Standard (PSM); (29 CFR
1910.119)
1994 - EPA List of Substances and Threshold
Quantities for accident prevention program
(40 CFR 68.130).
1996 - EPA Accidental Release Prevention
Requirements: Risk Management Program
(40 CFR 68).
Rather than impose new requirements for specific accident prevention measures, EPA
chose to rely in part on the public availability of RMPs, including the OCA information in RMPs,
6
-------
to help ensure that facilities take all reasonable steps to reduce their risk of accidental releases.
For many facilities covered by pre-existing accident prevention and response rules or voluntary
industry standards, the requirements to conduct an OCA and prepare a publicly available RMP
containing certain data elements from the OCA may be the only significant additional regulatory
requirements under the RMP rule.1 A complete description of the RMP elements, how an OCA
is conducted, and the various elements of OCA information contained in an RMP is available in
Appendix A.
The Agency decided that all RMPs would be submitted to EPA, which would handle
dissemination to state and local officials and the public. The Agency believed that this approach
would enhance dissemination and use of the RMP information. EPA's past experience in
implementing EPCRA had shown that many state and local officials needed assistance in
managing the chemical information submitted to them on paper by industry under that law, and
that the public often did not take advantage of this information since it was not conveniently
available.
With the help of the Accident Prevention Subcommittee2, EPA designed an RMP
reporting form that lent itself to the creation of an electronic database. The form consists of an
Executive Summary and sections for reporting OCA results, prevention program data, and other
information. In the Executive Summary, reporting facilities are required to explain in prose the
facility's risk management program, including a brief summary of the facility's OCA. The
remaining RMP sections, including the OCA sections, are in check-off box, yes/no and other
formats that allow compilation of an electronic database. As a result, the information in those
sections is relatively general in nature (e.g., the form calls for the facility to identify the types of
prevention devices it uses in a chemical process, but not where they are used or how many are
used). The vast majority of RMPs submitted by June 21,1999, were submitted electronically to
EPA. The Agency developed and maintains a central database of RMPs from which immediate
access can be provided to stakeholders who are designated recipients of the information as
mandated by Congress.
To satisfy the section 112(r) requirement that RMPs be made available to the public,
nearly all members of the Accident Prevention Subcommittee recommended that EPA place
1 The list and RMP rules are codified at 40 CFR Part 68.
2 The Accident Prevention Subcommittee to the Clean Air Act Advisory Committee was established to
provide the Chemical Emergency Preparedness and Prevention Office (CEPPO) with stakeholder advice and
counsel on scientific and technical aspects of its programs. The Subcommittee considers technical issues,
methodologies, and/or products which CEPPO provides for review. These form the basis for Subcommittee
findings and recommendations which enable CEPPO to strengthen its technical program and specific technical
products. The Subcommittee is made up of representatives from industry, state and local government, public
interest groups, academia, trade associations, and professional organizations. The Clean Air Act Advisory
Committee and the Accident Prevention Subcommittee were created under the Federal Advisory Committee Act.
7
-------
RMPs on the Internet for easy access by the public. However, concerns were raised that Internet
access to a large, searchable database of OCA results would provide a targeting tool for terrorists
and other criminals. In response, EPA decided not to place the OCA sections of RMPs on the
Internet, but concerns were next raised that recent amendments to the Freedom of Information
Act (FOIA) would compel EPA to release these sections in electronic format. Congress
responded by passing Chemical Safety Information, Site Security and Fuels Regulatory Relief Act
(CSISSFRRA), which the President signed on August 5, 1999.
CS1SSFRRA temporarily exempts OCA information from public disclosure under the
CAA and FOIA. It requires the President to assess the increased risk of terrorist and other
criminal activity associated with the posting of OCA information on the Internet, and the
incentives created by public disclosure of OCA information to reduce the risk of accidental
chemical releases. Based on the assessments, the President is to issue regulations governing the
distribution of OCA information in a manner that, in the opinion of the President, minimizes the
likelihood of accidental releases and any increased risk of terrorist activity associated with Internet
posting of OCA information and the likelihood of harm to public health and welfare.
The President delegated to the Department of Justice (DOJ) and the EPA authority to
perform the required assessments and to promulgate the required regulations. The President has
delegated authority to perform the assessment of the increased risk to DOJ, and has delegated
authority to perform the assessment of the incentives to reduce risk to EPA. The President also
jointly delegated to DOJ and EPA his duty to promulgate the regulations, subject to review and
approval by the Office of Management and Budget. (For a detailed description of CSISSFRRA
and the kinds of data and information available to the public and other stakeholders under
CSISSFRRA or by other means, see Appendix B.) This document reports the results of EPA's
assessment.
ORGANIZATION OF THIS REPORT
Because RMPs were submitted only in mid-1999 and because the vast majority of OCA
information is not currently available, it is not possible to analyze the impact of either complete
RMPs or OCA information directly at this time. EPA, therefore, has examined other programs
that provide the public with similar information related to risk. The assessment uses data from
these other programs as well as from the RMP program to answer a series of questions:
Chapter 2: Are Chemical Accidents a Serious Problem?
Before considering the risk reduction potential of public information, Chapter 2 addresses
the fundamental question of whether chemical accidents present a serious risk to the
public, employees of facilities, and the environment, and whether current regulatory
programs are in place to reduce the risk sufficiently. If chemical accidents pose little
threat because existing programs have already reduced risks, then there is less need or
benefit to be gained by making information available to the public for further risk
8
-------
reduction.
Chapter 3: Does Public Information Lead to Risk Reduction?
Chapter 3 examines the evidence that public information leads to risk reduction. This
question has two parts. Does the public use the data; and does that use lead to risk
reduction? In some cases, merely publicizing the data stimulates industry to take action.
Chapter 3 examines data from the Toxics Release Inventory (TRI) program, accidental
releases, and non-environmental programs.
Chapter 4: Does the Type of Information and Access Make a Difference?
Chapter 4 examines whether interpreted data, such as OCA data, are more likely to be
used by the public than raw data. Interpreted data are data that are easily understood by
the user without the need for further manipulation or supplemental information. Chapter 4
then examines whether ease of access to the data increases the likelihood that data will be
used by the public.
Chapter 5: Are There Other Sources of the Same Data?
This chapter examines whether OCA data can be obtained from other sources and whether
the availability is sufficient to lead to risk reduction.
Chapter 6: How Much Information is Necessary to Spark Risk Reduction Efforts?
This chapter examines ways OCA information could be made available and ways facilities
could be categorized or grouped by the hazards or risks of accidental release they present
in order to disseminate OCA information.
Chapter 7: What is the Public's Access to OCA Information under CSISSFRRA?
Chapter 7 describes OCA information in more detail, and
examines the options for data access under CSISSFRRA.
Chapter 8: Findings
Chapter 8 summarizes the findings and presents EPA's conclusions with respect to this
assessment.
In addition to the main body of the assessment, the report includes a number of appendices
for readers interested in additional details:
9
-------
Appendix A Provides a detailed description of the RMP data, including the OCA data and
discusses what is included in OCA data and what is not. A sample RMP is
included.
Appendix B Provides a detailed description of CSISSFRRA and its provisions that relate to this
assessment.
Appendix C Presents summaries of actions taken as the result of public environmental data.
Appendix D Presents details of a study of the effects of negative press on TRI emissions.
Appendix E Presents details of the accident data discussed in this report.
Appendix F Discusses the individuals and groups that are likely to use OCA information, and
how various uses can create or affect incentives for risk reduction.
10
-------
CHAPTER 2
ARE CHEMICAL ACCIDENTS A SIGNIFICANT PROBLEM?
Congress enacted section 112(r) of the CAA to reduce the number and severity of
accidental releases of chemicals that could cause serious harm. Although most localities are well
prepared for chemical emergencies, sudden accidental releases to the air that rapidly migrate off-
site (or, in the case of a flammable material, quickly reach an ignition source) potentially expose
the public or environment to harmful effects in a short time. As evidence of this, sheltering-in-
place has become a preferred emergency response strategy because air releases move too fast to
make evacuation a feasible option. Consequently, protection of public health depends on
preventing the releases.
Chemical accidents continue to be a serious problem in the U.S., causing deaths, injuries,
serious property damage, and disrupting business and the lives of individuals in the vicinity of
facilities. The facilities subject to the RMP rule submitted information on all of their serious
accidents that occurred in the five years prior to the date of submission of the RMP
(approximately June 1999).3 Serious accidents are defined in the rule as those that cause deaths
or injuries on- or off-site; significant property damage on-site; or known property damage,
evacuations or sheltering-in-place, or environmental damage off-site. Overall, 1,086 facilities
reported 1,913 accidents in their RMPs.
The impacts of the releases reported in the RMPs are shown in Table 1. It should be
noted that not all of the 1,913 accidents had one or more of the impacts that require reporting; a
few facilities chose to report all their releases of regulated substances rather than limit the reports
to those that were subject to reporting under the rule. No off-site deaths were reported in the
RMPs.
3 These data cover the five-year period prior to the submission of the RMP, but because each RMP has a
unique submission date, they may not cover exactly the same five years. However, most submitters probably
included accidents that occurred between mid-1994 and mid-1999. Therefore, the 1994 plus the 1999 numbers are
equivalent to one year's releases.
11
-------
Table 1 - Five-Year Accident History Data from RiMP Submissions
Year
Deaths
On-site
Injuries
On-site
Hospitalized
Other Medical
Treatment
Evacuated
Sheltered
in Place
Damage
($ millions)
1994
(partial)
6
239
46
135
3623
4,396
356
1995
2
433
103
4,823
8,677
21,978
67
1996
4
369
28
334
2,616
41,799
129
1997
5
416
11
583
7,267
65,041
218
1998
3
394
17
136
5,723
52,717
94
1999
(partial)
13
124
12
27
1,937
5,549
153
Total:
33
1,975
217
6,038
29,843
191,480
1,018
However, two recent accidents, one at a small chemical plant (4) and one at an ice-making
plant, have caused or contributed to off-site deaths. The ammonia release at the ice plant is the
first release of a toxic chemical in the U.S. known to have contributed to a fatality off-site as a
result of exposure to the chemical, rather than impact from an explosion (see box).
12
-------
Davie Ice Plant Leaks Ammonia - and a Grandfather Dies
Reprinted with permission from the Sun-Sentinel, Fort Lauderdale, Florida
By Tanya Weinberg
January 5, 2000
DAVIE — On the second to the last day of his life, Julio Lopez, 70, woke up and told his grandson how good
he was feeling. Early Wednesday, the last day of his life, he awakened to see his family panicking, while his
lungs bumed and a cloud of deadly ammonia hung low outside his home.
Something had gone wrong across the street at Reddy Ice, the ice-making plant on the 5000 block of 51st
Street. The pressure of a gas leak forced the top of a rooftop safety valve to pop off, and more than 100
pounds of ammonia was released, according to Reddy Ice. Because it was a humid, still night, the gas
absorbed moisture, sank, and invaded the lungs of workers and neighbors just after midnight, fire officials
said.
Two of seven employees were treated and released from Memorial Hospital Pembroke. Julio Lopez never
returned from Memorial Regional Hospital.
The cause of his death was heart disease and chronic emphysema, but inhaling ammonia is what "set him
off," said an investigator in the Broward County Medical Examiner's Office.
As his son led him running from their home, Lopez told him he couldn't breathe, that he couldn't go any
farther. He collapsed about a block from the house...
The cause of the leak and exactly how much ammonia was released have yet to be determined. Both the
Occupational Safety and Health Administration and environmental inspectors are investigating.
Table 2 assigns a dollar value to these impacts, based on the values EPA used in the
Economic Impact Analysis (EIA) for the final RMP rule in May 1996. Lost production is valued
conservatively, as it was in the EIA, at two times the value of property damage; according to
Marsh and McLennan4, the standard insurance industry assumption is that business losses are four
times the cost of property damage (5).
4 Marsh and McLennan represents the J&H Marsh and McLennan Corporation, parent company to M&M
Protection Consultants (M&MPC). M&MPC underwrites risk and provides consultation to management on hazard
control; they produce reviews of large property losses in the chemical and petrochemical industries.
13
-------
Table 2 - Dollar Values of Impacts of Releases in the RMP Five-Year Accident History
Impact
RMP Data
Unit Value5
Total
Deaths
33
$5,400,000
$ 178,200,000
Hospitalizations
217
$19,000
$4,123,000
Other medical treatment
6,038
$200
$1,207,600
Evacuation
29,843
$290
$8,654,470
Sheltered
191,480
$30
$5,744,400
Property Damage
$1,018,000,000
—
$1,018,000,000
1 net PrnHnrtinn
-------
standard is based on the concept that managing chemical accident risk requires an integrated
approach that involves identifying and assessing risks, managing risk through the adoption of
practices (such as operating procedures, training, preventive maintenance, management of change,
and periodic audits), and preparing for emergencies. This integrated system must be implemented
on an on-going basis. The RMP rule adopts the PSM standard as the basis for the prevention
program and streamlines it for facilities that pose lower levels of risk (based on accident history,
complexity of the process, and whether a worst case release could affect public receptors).
Many RMP facilities are also covered by PSM, yet continue to experience accidents.
OSHA does not require that facilities provide information outside the facility. Because there is no
list of facilities subject to OSHA PSM, it is not possible to determine with certainty whether
implementation of the PSM standard has reduced the number or severity of chemical accidents for
processes subject to the standard. Of the top seven sectors reporting accidents in the RMP
however, five (refineries, chemical wholesalers, chemical manufacturers, pulp and paper mills, and
cold storage facilities) are almost always subject to PSM for the same chemicals covered by the
RMP rule. The other two sectors (drinking water systems, wastewater treatment systems) are
covered by PSM in half the states. These sectors account for almost 80 percent of the accidents
reported in the RMP history of accidents since 1994. The largest number of releases reported in
the RMPs happened in 1997 and 1998 (1999 reports cover only a few months), and the number of
accidents reported in the RMPs for these sectors has remained fairly constant over the five-year
period. Even if PSM has reduced the accident rate, the number of accidents in these sectors
covered by PSM continues to be high.
CONCLUSION
Chemical accidents, despite previous regulatory efforts, continue to be a serious problem,
causing deaths, injuries, and property damage as well as public and business disruption.
Additional efforts are needed in order to improve chemical accident prevention practices.
15
-------
CHAPTER 3
DOES PUBLIC INFORMATION LEAD TO RISK REDUCTION?
A central question for this assessment is "Does disclosure of information to the public lead
to risk reduction?" Sharing information that affects the public has long had recognized value in
the United States. Congress, in a number of
laws, has provided people with access to
information that they can use to make better
decisions for themselves and society (see
box). This chapter discusses the evidence
that people will use information made
available to them and that the result of that
use is reduced risk. The next chapter will
consider whether the type of information and
the ease of access affect the level of use.
HAVE EPCRA DATA BEEN USED TO
REDUCE RISK?
In 1986, Congress passed the
Emergency Planning and Community Right-
to-Know Act (EPCRA) to improve local
planning for chemical emergencies and
provide the public with information about
hazardous chemicals in their communities.
Of relevance to this assessment are sections
312 and 313 of EPCRA.
• Section 312 requires facilities handling more than a threshold quantity (mostly 10,000
pounds) of certain hazardous chemicals to file an annual inventory with the state, Local
Emergency Planning Committee (LEPC), and the fire department. The inventory forms
are available to the public on request from the state or LEPC. More than 500,000
chemicals are covered by this section.
• Section 313 covers the Toxics Release Inventory (TRI) - annual reporting of releases and
transfers of about 650 chemicals. TRI reports are filed with EPA and the states and are
made available to the public on the Internet. Before the Internet was widely available, TRI
data were available through access to on-line data services (such as the National Library of
Medicine). The complete data set could also be obtained.
Section 312 and 313 data differ in some ways from OCA data. Section 312 and 313 data
Some Public Information Laws
and Programs
~ Freedom of Information Act
~ Federal Advisory Committee Act
~ Emergency Planning and Community
Right-to-Knovv Act
~ Clean Air Act §114(c)
~ Energy Policy and Conservation Act
~ Clean Water Act §308(b)
~ Resource Conservation and Recovery Act
§3007(b)
~ Food and Drug Administration processed
food labeling program
~ Securities Act of 1933, Securities Exchange
Act of 1934, and the Securities and
Exchange Commission's "EDGAR"
database
16
-------
are not interpreted to the extent OCA data are, and section 313 data represent actual releases
versus hypothetical releases. Despite these differences, using sections 312 and 313 data for
analysis is valid. This use is valid because EPCRA data are collected not only for emergency
planning, but also to ensure that the general public (as well as government agencies) is provided
with information about chemicals and chemical releases at facilities.
EPCRA section 312 data have not been widely requested or used by the general public.
The average number of requests for the data is one per year per LEPC or about 3,500 requests
per year nationally (6). Local emergency planners and responders do use the data (e.g., in the
1993 Midwest floods), but the public is generally not aware of the data. In a 1998 report, the
General Accounting Office agreed that section 312 data can be useful to a variety of groups, but
stated that the use of the data by the wider public has been limited. GAO suggested that costs
would be better justified if EPA improved the availability and use of the data by, for example,
ensuring that the data are computerized (7).
In contrast, TRJ data are maintained in a publicly available database and highlighted in an
annual report produced by EPA. From the first publication of the data in 1989, TRJ data have
drawn considerable attention from community and environmental groups, the press, state and
local governments, and industry. Environmental groups have used the data to publish their own
reports on specific areas or facilities (see box); the press has run numerous articles on local
facilities, highlighting those with high levels of emissions. Community groups have used the data
as the basis for actions. For example, in Northfield, Minnesota, two citizens groups and the
Amalgamated Clothing and Textile Workers Union joined forces and used safety and emissions
data to successfully bargain with Sheldahl, Inc. for a phase-out of toxic methylene chloride
solvent in favor of a non-toxic substitute, preventing a plant shutdown and loss of jobs. Research
shows that TRJ data are used by various groups to address environmental risk. Among public
interest and environmental groups, the three most frequently reported uses are directly pressuring
facilities for change, educating citizens, and lobbying for policy changes. State agencies that run
TRJ programs report their three largest uses are comparing TRJ data to permits, source reduction
efforts, and comparing emissions patterns at similar facilities. Industry's two most frequently
reported uses are source reduction efforts and educating citizens. Researchers have concluded
that TRJ data have been used to reduce environmental risk (8). In addition, TRI data have
spurred several states, such as Massachusetts and Louisiana, to adopt more stringent state
pollution laws (34, 35). Appendix C provides summaries of these and a number of other
17
-------
successful uses of TRI and other EPCRA
data.
HAS THE USE OF TRI DATA LED TO
RISK REDUCTION?
Since the first publication of TRJ data
in 1989, the emissions reported under TRI
have fallen about 43 percent (9).7 A number
of factors have caused the reduction. In some
cases, changes in the methods used to
estimate emissions have produced reductions.
Environmental groups have sometimes argued
that the reductions are mainly the result of
these different estimation methods. Industry
has argued that the reductions are real, driven
by the realization that product was being
wasted and opportunities existed to recover
valuable materials. Industry has sometimes
claimed that their emissions reductions are not stimulated primarily by the publication of
embarrassing data or the fear of public attention, yet an informal survey suggests otherwise.
A number of companies reduced their
TRI emissions to a far greater extent than the
general trend. One of the reasons these
companies may have made significant
reductions is due to negative press about their
environmental performance. This section and
Appendix D describe an informal assessment
whether media criticism and negative press
appears to have any positive effect on
subsequent toxic emissions (36). This
assessment was not intended to be an
exhaustive statistical analysis of this
hypothesis.
Although a facility may reduce its
toxic emissions for a number of reasons,
"manufacturers listed among the worst polluters ... may change their ways out of fear of customer
boycotts, increased regulation, or community hostility. The company's reputation, hard to build
7 Based on 1988 to 1997 TR.I data for core chemicals that have been subject to TRI requirements without
modification for the entire period.
Some Publications Using TRI Data
~ Toxics Hazards in Los Angeles (California
Public Interest Research Group)
~ Florida's Toxic Soup: Tracking Toxic
Trends
~ Toxic Air Pollution in Illinois (Citizens for
a Better Environment)
~ Clean Water Fund of North Carolina
~ West Virginia Scorecard (annual report by
the National Institute for Chemical Studies)
~ Poisons in our Neighborhoods (state reports,
Citizens Fund)
~ Poisoning the Great Lakes (Citizens Fund)
~ Richmond at Risk (Citizens for a Better
Environment)
~ Environmental Equity in Louisiana
impact of Public Information
in Indonesia
Indonesia's Environmental Impact Management
Agency implemented a pilot program, known as
"PROPER," in which certain industrial facilities
were given grades, based on their water
pollution performance. Researchers found that
public disclosure of these grades was sufficient
to prompt 10 factories to invest in pollution
abatement in order to improve their rating, and
lead to a more than 40 percent pollution
reduction in the pilot group in only 18 months
(10).
18
-------
and easy to destroy, is at stake" (33). Certainly, arbitrary actions to reduce emissions based on
fear of regulation or negative press without an understanding of risk can divert resources away
from situations that deserve greater attention and public concern. Consequently, it is even more
important to ensure that enough information is made public and that the public engage in dialogue
at the local level so that attention is focused on real risk reduction.
Several searches of newspapers, trade journals, and magazines revealed seven companies
that were repeatedly cited above others as "the worst polluters" in the nation, according to their
TRJ emissions, and several facilities that were repeatedly cited as "the worst polluters" in their
states. (See Appendix D for details.) Because of this, the TRI emissions data for each facility
were compared before and after it received negative press. In addition, for each of the nine
"worst polluting" facilities in states, a comparable facility located in the same region with the same
Standard Industry Classification (SIC) codes emitting the same chemicals as those identified as a
"worst polluter" was identified. However, these other facilities did not receive negative press for
their TRI emissions. Using the TRI database, EPA tracked the toxic releases for the "worst
polluting" facilities over the same time period as for the emissions from the comparable facilities.
The seven selected companies that were the subject of nationwide negative press about
their total toxic releases reduced their emissions and transfers 1.5 to 2.2 times more than the
general TRI trend in toxic releases and 1.3 to 19 times greater than the trends for their specific
industry sector as shown in Table 3.
Table 3 - Company-Wide Toxic Release Emissions Reductions 1990-1996
Company
Percent
Reduction
General
Trend
Industry
Trend
Improvement Over
General Trend
Improvement Over
Industry Trend
Inland Steel
95%
43%
5.2%
2.2 times
19 times
Kennecott
90%
43%
5.2%
2.1
17
Monsanto
84%
43%
51%
2.0
1.7
American Cyanamid
83%
43%
51%
1.9
1.6
IMC-Agrico
82%
43%
51%
1.9
1.6
DuPont
73%
43%
51%
1.7
1.4
3M _|
65%
43%
51%
1.5
1.3
At the facility level, the pattern is similar. Each of the facilities that was identified as the
subject of repeated negative local press reduced their emissions and transfers following the
negative press coverage. The comparable facilities did not. In one case, the facility reduced its
emissions of the chemical that was the subject of press coverage by 86 percent and its total facility
emissions by 64 percent. In contrast, all emissions and transfers for the other facilities owned by
19
-------
the company stayed relatively the same. (See
Appendix D for more details.)
Appendix C presents additional
evidence that companies respond to negative
publicity and public pressure by reducing
emissions and, therefore, reduce the risk to
the public and the environment. Other factors
certainly also produce emission reductions,
but facilities subject to public pressure do
appear to respond to that pressure and reduce
the risk that is subject to reporting. This
conclusion is further supported by the states
that have had chemical accident prevention
programs in place under state law since the
late 1980s and early 1990s. In Delaware and
Nevada, a number of facilities that were
complying with the state chemical accident
prevention laws switched chemicals or
reduced inventories to avoid having to file an
RMP with EPA that would become publicly
available.8 In both states, reports filed with the states are available only on request9 (14, 15).
Note also that with the growth in the economy, chemical use has increased. Industrial
production between 1991 and 1997 (the last year of TRI data) increased by 28.4 percent.
Chemical industry production increased by 19.5 percent. Consequently, the 43 percent average
reduction in TRI emissions and transfers reflects an even greater reduction per unit of production.
HAS THE PUBLIC FOCUS ON HAZARDOUS CHEMICALS REDUCED ACCIDENTS?
Accidental releases can be more difficult to prevent than routine releases (TRI covers
both) because the causes of accidents vary considerably; serious accidents are often the result of a
series of failures (human and equipment). A decline in the number or severity of accidental
releases would indicate that facilities have improved practices generally.
8 Other explanations for why these facilities exited the Delaware and Nevada programs are unlikely. For
example, the sources that exited these programs are likely to have already incurred most of their compliance costs
due to the similarity of EPA's and the state's requirements (other than disclosure).
9 New Jersey and California also have state programs. More than two thirds of the original registrants in
New Jersey switched chemicals or reduced inventories to avoid the state law, which involves large annual fees
($10,000 to $100,000). It is unlikely that those that remain in the program have the option of avoiding the federal
program. California delegated its program to 147 local agencies and keeps no central records so that trends in that
state cannot be traced. California also uses much lower chemical thresholds than the federal program.
Publicity Brings Risk Reduction
The power of press attention to produce action
was seen in Washington, D.C. in late 1999,
when the Washington Post ran a series of
articles on concerns for chemical safety at the
Blue Plains Wastewater Treatment Plant. The
articles featured scenario information that the
plant (voluntarily) included in its RMP
Executive Summary, the health effects of the
chlorine and sulfur dioxide used at the plant, and
apparent problems with plant safeguards. As a
result of these articles, the D.C. Water and
Sewer Authority and the mayor immediately
committed to resolving all of the safety problems
cited in the articles. The Authority has
improved security and now plans to replace its
chlorine one and a half years sooner than
originally planned (11)(12)(13).
20
-------
EPA examined whether certain hazardous substance releases have declined in the 1990's.
As noted in Chapter 2, facilities are required to report episodic releases of hazardous substances
that are above a reportable quantity (RQ) to the federal government. Data from these reports are
compiled in EPA's Emergency Response Notification System (ERNS) database. ERNS also
includes petroleum product spills (crude oil, processed oil, diesel, jet fuel, kerosene, gasoline); the
database covers both transportation and fixed facilities. From a high of 7,800 in 1994, hazardous
substance releases recorded in the ERNS database declined to 5,400 in 1999.
Because ERNS data include many reports of unknown material or unknown sources, and
for hazardous substances, many reports of substances that are not subject to reporting and of
quantities below the reportable quantities, EPA narrowed the analysis to four states,
Massachusetts (MA), Connecticut (CT), Virginia (VA), and New Jersey (NJ). Focusing on four
states allowed for a more detailed analysis of the data. These were selected because the four were
relatively similar in size and industrial sectors and include a broad range of industries. The
analysis looked at all ERNS-reportable hazardous substance releases from fixed facilities,
excluding those generated by private citizens or unknown sources and those of unknown
materials. The analysis also looked at TRI chemicals reported by manufacturers, and finally, all
reportable releases based on the current RQs for currently listed substances. Because the
reportable quantities and chemicals listed changed in the early to mid 1990s, EPA made sure to
use consistent criteria when comparing reported releases.
The results of all the analyses were similar, as can be seen in Exhibit 1. In these four
states, all hazardous substance releases in ERNS from fixed facilities declined by 60 percent from
their peak year (1992). Similarly, episodic releases of TRI chemicals from manufacturers (the
only group subject to TRI) and releases of hazardous substances above current RQs declined by
about 68 percent from their peak year (1990). Because the analysis considered only currently
listed substances at current RQs, the 1995 RQ adjustments are not responsible for the decline.
(Note that these numbers cannot be compared to national hazardous substance numbers, which
include a substantial number of releases from unknown sources and of unknown materials, as well
as some releases generated by private citizens.) Appendix E provides the data from the four states.
21
-------
Exhibit 1 - Episodic Hazardous Substance Releases - Four States
CO
z
QL
LU
C
0)
-------
Exhibit 2 - Oil and Transportation Releases - Four States
2500
2000
"¦1500
(D
0)
0)
CL
o
a>
_o
| 1000
2
500
97
9fi
92
99
91
90
Year
—U Oil —*— Transportation —*— Average oJ Oil and Transport
has occurred for substances and modes not subject to the same level of public scrutiny.
IS THERE EVIDENCE THAT OCA INFORMATION AVAILABILITY WILL LEAD TO
PUBLIC ACTION?
Because RMPs were filed just last year and the public has had little or no access to OCA
information, there has been little opportunity for the public to use RMP or OCA data to pressure
industry to reduce risks. The chemical industry, however, has conducted pilot projects to provide
the public with OCA data. In 1994 in the Kanawha Valley of West Virginia, industry, in
conjunction with local emergency planners, developed worst-case release scenario data and held a
public meeting to inform citizens about the risks posed by hazardous chemicals in their
community. The event was held in a local shopping center that citizens frequented on Saturdays.
Booths were set up and information was distributed that explained the chemicals used by local
industries and what health effects could occur if those chemicals were accidentally released.
Armed with this information, citizens since have formed groups and have gone to companies to
persuade them to discuss their overall environmental management strategy, including using less
toxic substances. Additionally, other citizens worked with industry officials to prepare and
distribute a kit to community residents that included instructions and materials that could be used
to shelter-in-place during a chemical emergency (16).
23
-------
Such pilot projects have been led by the chemical industry, which has made outreach a
part of its Responsible Care® program. These voluntary efforts, though laudable, are rare, and
have generally not included facilities in other sectors. The chemical industry and refineries
represent only about 13 percent of RMP facilities. Thus, except for a few locales, the outreach
about accident scenarios that is needed to promote public involvement has not occurred.
IS THERE OTHER EVIDENCE ON THE IMPACT OF PUBLIC INFORMATION?
Although other programs outside of the environmental area are not directly comparable to
the OCA situation, the success with these programs none-the-less illustrate the value of public
information. For example, a 1981 National Highway Traffic Safety Administration study found
that, after 1976, purchasers of new automobiles in the U.S. actually used fuel economy test data
to help choose fuel-efficient vehicles (17). A 1985 study made a similar finding (18).
The Food and Drug Administration (FDA) has sought to improve public health via
product labeling requirements for a number of years. For example, comprehensive changes were
made to processed food labels in 1994. Research shows that these enhancements to food
nutrition labeling have, in combination with other factors, produced health benefits. Studies show
that not only do consumers read nutrition labels more often, but they also reduced the fat in their
diets (19). Now, based in part on the positive effects of the 1994 label revisions, the FDA has
recently proposed adding label requirements for trans-fatty acids, to further reduce rates of heart
disease and early death.
There are many ways in which access to OCA information could yield positive impacts,
based on incentives created or affected among various segments of the public. These incentives
and the likely uses of OCA information by the public are detailed in Appendix F.
CONCLUSION
The public (which includes the media, public interest groups, as well as citizens) does use
available information on risks and the availability of such information correlates with a substantial
reduction in the risks that are the subject of reporting.
24
-------
CHAPTER 4
DOES THE TYPE OF INFORMATION AND TYPE OF ACCESS MAKE A
DIFFERENCE?
The first part of this chapter addresses two questions: Does the type of information
available influence the degree of risk reduction caused by its availability? Does the type of access
to the information affect the use of the data?
A. DOES THE TYPE OF INFORMATION INFLUENCE DATA USE?
RMP data other than OCA data are already available on the Internet through RMP*Info10.
The RMP data contain information that could be used to generate off-site consequence results,
although these would be less accurate than the OCA data reported by facilities (see Chapter 5).
OCA data are potentially valuable because they are "interpreted" data. For example, rather than
simply providing the public with raw numbers for the quantities of chemicals on-site, the OCA
data provide the public with information about what might be affected if a worst-case or
alternative release involving that quantity should occur. The public does not need to do anything
further with the data to understand why it should be concerned (or not) about chemicals at the
facility.
TRI data are partially interpreted data. They tell the public how much of a toxic chemical
is being released into the community and the media (air, water, etc.) to which it is released. But
unless the public is familiar with the hazards of the chemical and industry practice for handling
that chemical, the public may not understand whether the emissions pose a risk and whether those
risks are high or low in comparison to other facilities in a locality or across the U.S. TRI data
have always been available electronically. The data were originally posted in the on-line database
of the National Library of Medicine and could be searched both by facility and on combinations of
other parameters. TRI data are now available in at least three Internet databases:
• EPA's Envirofacts database (www.epa.gov/enviro), which presents some of the data with
simple queries; more advanced queries are possible, but not simple to program. No
interpretation is provided.
• RTK-Net, which provides access to all data on each TRI form, without interpretation.
RTK-Net (www.rtknet.org) also provides access to multiple other environmental
databases.
10 RMP*Info™ is the name of EPA's database of RMPs submitted by facilities. RMP*Info™ is located in
the Envirofacts Warehouse on the Internet at http://www.epa.gov:9966/srmpdcd/owa/overview$.startup.
25
-------
• Scorecard, the database from Environmental Defense (formerly Environmental Defense
Fund), which was designed for easy use and provides interpretations of TRI and other
environmental data on each facility. For each facility, Scorecard (www.scorecard.org)
provides not only chemical names, but the hazards of the chemicals, how the facility ranks
in the U.S., state, and county on a number of measures (e.g., total releases, cancer risk,
non-cancer risk), quantity of releases aggregated by hazard, estimated cancer risk, and the
facility release history for each year since 1987. The site also allows a user to map the
facility in the county and to send a free fax to the facility.
Public reaction to the Internet launch of Scorecard in April 1998 is a measure of the value
of interpreted data. At that time, RTK-Net was registering about 240,000 searches of its
databases a year. Scorecard has registered 2 million visitors since its launch and serves about
600,000 page views" a month. Site visitors have sent about 5,000 faxes to about 3,000 distinct
companies (20). By providing more highly interpreted data, Scorecard has increased the
accessing of data on the Internet by a factor of up to 30. As a point of comparison, from July
1999 through January 2000, without OCA information, EPA's RMP*Info has logged over
155,000 page views of RMPs or about 23,000 per month (or about four percent of Scorecard's
page views). A comparison with Scorecard is not entirely fair; Scorecard provides other
environmental data and ease of use, while RMP*Info is new, concerns only accidental releases,
and does not contain OCA information.
Another reflection of the value of interpreted data is that in the early years of TRI, few
reporters used the TRJ database directly (only 20 percent in one review of large dailies); the large
majority waited until the government or public interest groups published interpreted data (21).
Overall, evidence strongly supports the argument that the form of the data does affect the degree
to which the public will use it. Interpreted data, such as OCA data, will be used far more than
raw data.
DOES EASE OF ACCESS TO THE DATA AFFECT USE?
As mentioned in the previous chapter, EPCRA section 312 requires thousands of facilities
to file annual forms reporting on the quantities of hazardous chemicals they use or store. Section
312 covers all hazardous chemicals (estimated to number about 500,000) and almost all industrial
sectors. The 312 forms, therefore, may be the most complete record of hazardous chemical use in
the U.S.
These forms are filed with local emergency planning committees (LEPCs), state
emergency response commissions (SERCs), and local fire departments. LEPCs and SERCs are
required to make the forms available to the public, but, because of limited resources, only about
'1 A "page view" is the downloading of an HTML page. The number of Scorecard page views would tend
to be somewhat higher than the number of database searches for the same activity, because most but not all of its
page views pull data from a database. "Hits" are not used here because they are less accurate indicators of data
access.
26
-------
1000000
Exhibit 3 - Impact of Access and Type of Data
B00000
| 600000
O
CD
(0
CL
o
§ 400000
O"
0)
0C
200000
Q_J , . 1 , 1 1 ,
EPACRA 312 Data via LEPCs RTK-Net Scorecard
half advertise the forms' availability. If a member of the public wishes to see the forms, he or she
must locate the SERC or LEPC and request the information. According to a survey conducted
for EPA by George Washington University, the average number of requests for these forms is one
per LEPC per year (22). The national total, therefore, is about 3,500 requests per year. In short,
obtaining section 312 data requires an effort to seek it out; few people have made the effort. A
further disincentive, as noted above, is that the data are often not laid out in a format from which
the public can easily gain value.
The results are similar for state chemical accident prevention programs. As discussed in
Chapter 3, New Jersey, Delaware, and Nevada have had accident prevention program rules in
place since the late 1980s and early 1990s. In each state, facilities file reports, but these are not
made available to the public except on request. New Jersey reports that since its program began
in 1988, less than 10 individuals or organizations have requested facility reports or information
(23). Delaware has received similarly few requests (24). Nevada places its reports in public
libraries, but its experience indicates that librarians have to hunt for the files when updates are
needed. Few people have requested information from the state unless there has been an
accident(25).
This lack of interest and effort contrasts sharply with the number of people seeking data
from RTK-Net and Scorecard, where the data can be accessed at any time from office or home.
Comparing the 3,500 requests a year via LEPCs to the 240,000 database searches per year for
Internet-based RTK-Net is consistent with the finding that providing data on the Internet has
increased the use of data many times over. Comparing the LEPC request number to Scorecard
visitors per year yields a difference of a factor of more than 280.
27
-------
CSISSFRRA requires facilities to hold a public meeting or post a notice summarizing their
OCA information. EPA has not surveyed facilities to determine attendance at these meetings;
however, a very limited number of facilities have provided feedback. One large California public
agency held seven meetings in different parts of Los Angeles, Orange, and Riverside counties.
The average attendance was two to five, though at the final meeting, held on a Friday night, one
person came. For each meeting the agency ran an ad twice in the local paper and sent 500 to
1,000 direct mail announcements to households within one mile of the facility. Other reports have
been similar, though one meeting covering several facilities had 500 people attend. That meeting
was in Henderson, NV, the site of two major accidents in the last 11 years. In contrast, when
industry has gone out to the public (e.g., at a shopping mall in Charleston, West Virginia) rather
than making the public come to industry, outreach and communication about chemical accident
risks have been much more successful.
CONCLUSION
The public will make use of data and information to the extent the information is
understandable, useful, and easy to use. The more effort is required on their part, the less likely
they are to seek out the data. If data are easy to access, use will rise significantly. If data are
interpreted rather than raw, use will expand even more. Exhibit 3 shows the numbers of
requesters for section 312 data sought from LEPCs, raw data from RTK-Net, and interpreted
data on Scorecard.
28
-------
CHAPTER 5
ARE THERE OTHER SOURCES OF THE SAME DATA?
Another question that must be considered is whether the public already has access to OCA
data or can easily derive such data without the dissemination of the OCA sections of RMPs. This
chapter considers what is currently available and whether this information is an adequate
substitute for the data submitted as part of the RMP.
WHAT OCA OR OCA-RELATED DATA ARE AVAILABLE TO THE PUBLIC NOW?
One potential source of OCA data is the RMP Executive Summary. By regulation, each
RMP's Executive Summary must include a brief description of the worst-case release scenario(s)
and the alternative release scenarios(s), including administrative controls and mitigation measures.
These Executive Summaries are posted on the Internet. However, neither regulations nor
guidelines from EPA specify how much information should be in the Executive Summaries. They
vary greatly in the amount of OCA data they provide. In addition, they are in narrative, or text,
format, so they cannot be easily searched on variables of interest. Executive Summaries may or
may not provide information of interest regarding a particular facility, and they cannot be relied
upon to provide a full picture of a group of facilities or a basis for comparing facilities.
Other publicly available data can be gathered and analyzed to provide information similar
to some of the OCA data. For example, important components of the worst-case release scenario
are the identity and the quantity of toxic or flammable substance held in the largest vessel at the
site. The chemical name and quantity of that chemical in each covered process is available in the
registration section of the RMP. However, for a large facility with several listed substances or
covered processes, the public will not know which chemical or process was considered for the
potential worst-case or alternative release or how much of the chemical was expected to be
released. Even if only a single chemical is at the site, the quantity of the chemical in the process
may not be the same as the quantity stored in the largest vessel.12 Storage quantity is also
available through TRI and other Internet or publicly available databases, but these data reflect
total quantity on site, reported in broad ranges. If the quantity of the regulated substance listed in
the RMP or provided by TRI is used, it is likely that the scenario would over-estimate the
consequence distance since the maximum quantity in a process or on-site is being used in lieu of
the maximum amount stored in a single vessel.
For the worst-case scenario, the public has access to an EPA model or software tool
which most facilities have used to perform their analyses; however, without certain key inputs
(quantity in the vessel and specific chemical), people may get results that are very different (and
12 In some cases, the largest vessel may not be the source of the worst-case release if a release from
another vessel closer to the public would have greater off-site impacts.
29
-------
potentially more alarming) than those provided by the facility. For alternative scenarios, the
facility has wide latitude in its choice of inputs; without these same inputs, a member of the public
has no way to reproduce a facility's OCA data.
The potentially affected population is a function of the distance that a toxic cloud travels
or that an explosion affects, so whatever errors are made in determining this distance will also
affect the public's calculation of population affected.
In contrast to information dependent on the release distance, more general siting and
surrounding population data is more easily accessible. Census and other data for determining the
population of a given area are publicly available. Facility location information can easily be found
in telephone directories (e.g., yellow pages). In particular, sources of the yellow pages on the
Internet are also linked to a map that shows the exact location of facilities. Other siting
information for major facilities of publicly traded companies appear in 10-K reports available on
the Internet through the Securities and Exchange Commission's Electronic Data Gathering,
Analysis, and Retrieval (EDGAR) system.
The prevention program portion of the RMP lists all the types of mitigation measures used
by a facility in a process, but without OCA-specific data the public will not know which measures
should be involved in calculation of consequence distance or the extent to which the measure
might mitigate a release.
HOW EASY WOULD IT BE TO REPLICATE AN OFF-SITE CONSEQUENCE
ANALYSIS?
Data are available to conduct an OCA, even if the results are likely to differ considerably
from those submitted by facilities. How easy is this process for a citizen?
Accessing and reading the regulations and the guidance would be an initial step, followed
by acquiring the mathematical models needed to turn quantity data from the RMP or other
sources into concentration estimates around the facility. Information on the locality around the
facility (population, land-forms, buildings and other structures, environment, meteorology, etc.)
would need to be acquired. Information peculiar to the facilities would need to be gleaned from
the raw data in the RMP, if, indeed it was actually available. Finally, the user would have to
integrate this information and tools, check the results, and produce estimates. Although members
of the public are not likely to perform such tasks, determined individuals or organization(s) could
do so relatively easily using information readily available.
Clearly additional economic costs to society for producing such estimates by outside
parties would be incurred. The level of costs would depend on how many such independent
assessments might be made, and how many of these assessments might be made by different
parties; it would be impossible to estimate with any certainty. However, it seems very likely that
the summation of the costs of outside assessments would be considerably greater than the costs
that have already been incurred by the facilities in producing one OCA per facility. These are
30
-------
sunk costs, (i.e., already incurred), and were estimated at the time of the RMP rulemaking at $44
million, nationwide. OCA-specific information already exists. Any re-estimation by outside
parties would entail additional societal costs beyond those which have already been expended.
From the point of view of economic efficiency, having the OCA analysis produced once (by the
facilities) is far preferable to having outside users attempt to do so themselves.
DOES THE OCA INFORMATION PREDICT ACTUAL RELEASES OR ACTUAL
CONSEQUENCES?
As described in more detail in Appendix A, the OCA information represents hypothetical
estimates of the consequences of worst-case (WCS) and alternative release scenarios (ARS).
They do not predict likely releases or consequences. Worst-case scenarios in particular are based
on extreme conditions. Many of these conditions cannot be controlled or predicted. Some of
these conditions and a description of their characteristics are:
• Weather. The WCS is evaluated at very low wind speeds and very stable atmospheric
conditions like what might occur at dawn on a warm summer morning. A smoke plume
from a fire-place might look like a very lazy ribbon as it travels slowly downwind (due to
less dispersion). This condition occurs fairly infrequently (26); most of the time there is a
slight breeze and some instability in the atmosphere which acts to quickly disperse a toxic
cloud. A worse condition that sometimes occurs, especially in the summer, is an
atmospheric inversion which can magnify pollution effects. The ARS is evaluated at more
"typical" weather conditions - moderate winds and atmospheric stability. No one can
control the weather conditions at the time of an accidental release.
• Affected population and wind direction. The OCA information contains estimates of
affected populations inside worst-case and alternative release scenario circles whose radius
is equal to the downwind distance the cloud travels to a toxic endpoint (see below). Toxic
gas clouds generally travel in the direction of the prevailing wind; they form a long,
narrow plume, which covers only a relatively small fraction of the worst-case circle. (See
Exhibit 4.) Therefore, depending on wind direction, only a small fraction of the people
reported in the OCA information would actually be affected by a cloud. In a sense, the
population reported is a sum of the people under all worst-case plumes emanating from a
location; any individual plume will cover far less than the sum. For flammable gas
31
-------
scenarios, the blast effects would be felt in all directions from the source, so all people
inside the circle could feel the effects, depending on the characteristics of the blast.
Exhibit 4 - Typical Plume Map (27)
filOYClOWJl
fMCWV' *
Gcrtfcn
Actual Exposure or Endpoint. As mentioned above, the OCA information includes an
estimate of the population affected by a WCS or ARS within a circle whose radius is
determined by the distance from a point of the release to a toxic or flammable endpoint.
This number is not an estimate of the number of fatalities or injuries that would occur
following an actual release, because the endpoint is not the concentration where there is a
32
-------
100% likelihood that an exposed person would die or suffer injuries. The concentration
that would result in death, for example, is much greater and would be much closer to the
source of the release. The population number represents the people that could potentially
be exposed; actual exposure and the effects of exposure would depend on where a person
is (indoors or outdoors) and for how long he or she remains in the gas cloud, and
breathing the chemical. For flammable substances the endpoint for a worst-case explosion
is only one pound per square inch (psi) over-pressure; fatality levels are much higher (ear-
drum rupture is 2.3 psi).
• Amount Released. The WCS assumes immediate (or within 10 minutes) release of the
maximum quantity of chemical in the largest storage vessel. However, there is no
guarantee that a vessel will always hold the maximum amount; this may occur
infrequently. In addition, chemicals flow out or vaporize from failed vessels according to
the laws of physics; there may be an initial surge followed by a gradual reduction in flow,
taking more time than that given by the WCS. Modeling these conditions is complicated;
the WCS is designed to be a simple, straight-forward estimate.
Studies of severe chemical plant accidents have shown that such accidents have usually
resulted from the confluence of multiple abnormal events or conditions in process or management
systems along with unusual meteorological conditions over which no one has control (28). For
example, in Bhopal, four separate safety systems, any one of which would have prevented the
accident, had been disabled prior to the accident, and a fifth failed to operate properly (29).
Further, weather conditions were such that the chemical cloud did not disperse, exposing more
people than if it had dispersed.
In the case of flammable materials, an initiating explosion will most often immediately
ignite any flammable material released, causing a large fire, but possibly preventing a much more
severe vapor cloud explosion (the worst-case flammable accident). Vapor cloud explosions
require that flammable gas be released into a somewhat confined area for a period of time prior to
gas ignition. If the gas is immediately ignited before sufficiently large quantity is released, a fire
usually occurs, but generally not an explosion. Numerous experimental programs devoted to the
study of vapor cloud explosions have shown that such explosions are difficult to reproduce, even
under carefully controlled conditions (30). Even accidental explosions at facilities that store,
transport, and manufacture explosives have usually resulted in little off-site damage (31).
CONCLUSION
Although the public could educate itself and attempt to reproduce OCA data from other
sources, the effort involved is likely to discourage the vast majority of people. At the same time,
some OCA data are available in RMP Executive Summaries, which are available on the Internet.
Actual chemical releases are different from the releases evaluated in the OCA information.
No one can control all of the conditions (for example, weather) used to develop the OCA
information; actual conditions at a facility can vary widely from those used in the OCA
33
-------
assessment. However, facilities subject to the RMP requirements generate the OCA information
according to consistent guidelines so that the public and others can understand the relative
potential hazards and risks present at a facility.
34
-------
CHAPTER 6
HOW MUCH INFORMATION IS NECESSARY TO SPARK RISK
REDUCTION EFFORTS?
The previous chapters describe how OCA data might be used by a variety of stakeholders
for chemical accident risk reduction. In order for OCA data to be most effective as an incentive
for risk reduction, data from a sufficient number of facilities must be disclosed to the public.
Disclosure of a sufficient number of facilities' OCAs is necessary to identify plausible relationships
between the hazards posed by facilities and their accident prevention and mitigation actions and
the risks of future chemical accidents. A plausible relationship can be determined from
comparisons between similar facilities (or processes) and chemicals. This chapter will provide
several options for the public to obtain comparison information based on OCA data as well as the
number of RMPs submitted that would support each option.
WHAT INFORMATION IS IMPORTANT FOR RISK REDUCTION?
Risk is a measure of the potential for economic, environmental, or human loss. Chemical
accident risks are generally estimated as a product of the quantity and toxic or flammable
characteristics (chemical hazards) of a substance, the population potentially exposed by the
release, and the likelihood that the chemical will be accidentally released. For example, if a small
release of hazardous chemical occurs frequently (indicating that it is likely), but the chemical does
not generally migrate off-site, the overall risk to the public is probably low. If the likelihood of a
catastrophic release of a large quantity is extremely low, but the number of people affected if it
did occur is large, the overall risk may still be low because of the low probability of release. If a
large release could occur relatively frequently affecting a large number of people, the overall risk
to the public is high.
All else being equal, risk reduction occurs when either the chemical hazard, population
potentially exposed, or likelihood are reduced. The toxic or flammable characteristics of a
chemical generally cannot be changed; for example, chlorine is toxic when inhaled, propane is
flammable. However, reducing the quantity of hazardous chemicals on-site or switching to less
hazardous substances will reduce risk. The population potentially exposed can be reduced by
minimizing the amount available to be accidentally released or by mitigation measures that capture
or destroy an accidentally released substance before it travels off-site. A relative sense of the
likelihood for accidental releases at a particular facility can be gained from its accident history;
reducing the likelihood of accidental releases is the main goal of the RMP program.
RMP and OCA data give the public some insight into all of these factors. However, the
most straightforward ways the public is likely to measure whether risk is being reduced are a
reduction in the number and severity of accidental releases and population potentially exposed.
The distance to an endpoint and the estimated population and public receptors within this distance
35
-------
serve as indicators of the potential size and impact of an accidental release at a facility.
HOW IS INFORMATION USED TO ENCOURAGE ADDITIONAL RISK
REDUCTION?
Risk reduction begins when an individual who can initiate actions for risk reduction (e.g.,
facility management, a community member, local media) believes that the current level of risk is
unacceptable. Dialogue at the local level among the public, the LEPC, first responders, and the
facility is the best way to raise concerns and to discuss possible options.
RMP and OCA information from a single facility provide insight about risk that could
affect an individual. But by itself, developing a clear understanding of risk and whether it is
acceptable is difficult. Since everything we do is "risky" to some extent, risk-based decisions
must rely on a comparison with other familiar risks or known benchmarks; consequently a sense
of the risk posed by one facility is best established by comparing it to other facilities. From this
comparison and understanding of potential risk, unacceptable risks can be reduced through a
reduction in hazards (e.g., lower quantities of chemicals on site) or likelihood (e.g., improving
accident prevention programs or mitigation systems) or both.
One simple approach to comparing the hazards posed by facilities is to compare distances
to an endpoint. An individual can compare the distances for multiple facilities and obtain a
relative ranking of the maximum and more likely extent of risk posed by the facility. Distances to
an endpoint can be compared because the distance to endpoint represents a similar level of
potential impact. Likewise, the residential population and the public and environmental receptors
within the distance to endpoint provide additional insight into the potential extent of the damage
associated with a release and, hence, an aspect of risk. However, two facilities could have the
same endpoint distance but at one facility no public receptors are within this zone. These
approaches, however, do not convey information about the likelihood of a release. For example,
given two facilities with the same distance to endpoint or similar number of public receptors, one
could have a much higher likelihood of a release, and thus a higher risk.
HOW MANY RMPs OR OCAs ARE NECESSARY FOR FACILITY COMPARISONS?
A. Comparisons across facilities in the same locality
An individual may want to compare the distances to an endpoint for a particular facility
with distances for all other facilities within a given locality, for example, a county, or, if close to a
county border, two or more counties. (Most LEPCs are county-based, and a number of facilities
have distances to endpoint of 25 miles.) Table 4 shows the counties with the most RMP
facilities. Individuals within these counties would need access to at least this number of complete
RMPs (including OCA information) to develop county-wide comparisons. Alternatively, an
individual could collect the RMPs within a state, city, or zip code; but because of wide variability
in the number of facilities within states, cities, or zip codes and since most LEPCs are established
by county, a county-wide comparison seems to make the most sense.
36
-------
Table 4 - Counties with the Most Facilities Submitting an RMP
County:
Number of RMP Facilities:
Harris County, TX
201
Los Angeles County, CA
162
Kern County, CA
96
Cook County, IL
76
Maricopa County, AZ
71
Table 5 shows how the number of RMP submitting facilities are distributed across
counties. The RMP database has RMPs from 1,501 to 2,384 counties in the US; these counties
have at least one RMP up to 201 RMP facilities (Harris County, TX). Ideally, residents should be
able to see all the RMPs and OCA information for the facilities in their county; if they are
restricted to 10 RMPs for example, residents in 82% of the counties that have RMP facilities
would be able to review all RMPs in their county while 18% would be unable to do so.
Table 5 - Distribution of RMPs and Counties
Number of
RMP Facilities
Number of
Counties
Portion of all Counties that
have RMP Facilities (percent)
1 - 5
1501
63%
1 - 10
1955
82%
1 - 15
2222
93%
1 - 20
2309
97%
1 - 50
2376
99%
1 - 100
2383
99%
1 -201
2384
100%
But comparing facilities based on distance to an endpoint within a county is not likely to
be enough. The public needs to know more to be able to compare facilities and form an
understanding of the risk. For example:
• Is the quantity of chemical stored reasonable?
• Are there mitigation or other measures that could be used to reduce the distance to
endpoint?
To answer these questions, the public needs to be able to compare RMPs for similar facilities
using the same chemicals.
37
-------
B. Comparisons based upon Type of Process
The best way to compare a process at one facility with other similar processes at other
facilities is to choose processes that are categorized in the same industry sector. When a facility
prepares an RMP, processes are assigned a code using the North American Industry Classification
System (NAICS). The NAICS codes identify the industry segment for that process. For
example, Agricultural Supply processes have a NAICS code of 42291. If someone wants to
compare a process with all others in its industry sector, then all RMPs that have a process in the
same industry sector would be needed.13 Table 6 lists the top twelve industry sectors (based on
NAICS code) that had the greatest number of submitted RMP processes. For example, if
someone wants to compare OCA data from an agricultural supply process (NAICS 42291) in
their area to all other agricultural supply processes, then the OCA data from 4,033 other
agricultural supply RMP processes would be needed.
Table 6 - Industry Sectors Submitting the Most RMPs
Industry Sector Description - NAICS Code
RMP
Processes
Percent of All
RMP Processes
Agricultural supply - 42291
4,034
27
Water Treatment - 22131
1,892
13
Sewage Treatment - 22132
1,361
9
Refrigerated Warehouses - 49312
504
3
Natural Gas Liquid Extraction - 21 1 1 12
450
3
Other Chemical and Allied Products - 42269
356
2
Farm Product Warehousing - 49313
326
2
Liquified Petroleum Gas Dealers - 454312
307
2
Support Activities for Crop Production - 1 151 1
283
2
Plastics Material and Resin Production - 325211
250
2
Basic Organic Chemical Manufacturing (Not otherwise specified) - 325199
244
2
Poultry Processing - 311615
216
1
13 EPA has prepared "Model RMPs," for certain industries such as chemical distributors,
warehouses, ammonia refrigeration, and sewage (wastewater) treatment. Typical processes in these
industries are very similar making standardized accident prevention programs possible. The OCA
information for a process in one of these industries could be compared to the model. However, the model
does not characterize a typical facility nor would this provide a comparison of distances to endpoint
submitted by other sources in that industry.
38
-------
Table 7 shows how the reported industry sectors are distributed across all RMP
processes. For example, 315 industry sectors (or 67% of the total number of industry sectors
reported in the RMP database) have 5 or fewer RMP reporting processes within that sector.
Likewise, 452 industry sectors (or 92% of the total) have 50 or fewer RMP reporting processes
within that sector. As above, using the data in Table 7, if the public had access to the OCA
information in 10 RMP processes, then comparisons could be made across 76% of all industry
sectors. However, this would not be enough access to allow full industry comparisons for 113
industry sectors or 14% of the industry sectors.
Table 7 - Distribution of RMPs Across NAICS Codes
Number of
RMP Processes
Number of Industry Sectors (NAICS
codes) having this number of processes.
Percentage of all NAICS codes having
this number of processes
1 - 5
315
67%
1 - 10
362
76%
1 - 15
389
82%
1 - 20
403
85%
1 - 50
435
92%
1- 100
452
95%
1-4,034
475
100%
C. Comparisons by Specific Chemical
Another way to compare facilities and processes is to compare the OCA results for the
same chemical across different facilities. For example, someone may want to compare the OCA
data for a local facility that uses ammonia to all other facilities that use ammonia. Or, someone
may want to examine the OCA data for facilities using the same chemical that had an accidental
release in their five-year accident history. Table 8 lists the chemicals most often reported in
worst-case scenarios and in the RMP five-year accident history. According to this Table,
ammonia OCA data from 7,506 other facilities would need to be accessed. Similarly, if someone
wanted to compare the OCA data for facilities that had accidents involving ammonia, then
ammonia OCA data from more than 600 other facilities would need to be accessed.
39
-------
Table 8 - Chemicals Most Often Reported in RMP WCS and 5-Year Accident History
Chemical
Number of
WCS Reported
Percent of
all WCSs
Number of
Accidents
Percent of
accident history
Ammonia
7506
45
656
34
Chlorine
4104
24
518
27
Propane
1228
7
54
3
Flammable Mixtures
847
5
99
5
Sulfur Dioxide
445
3
48
3
Ammonia (aqueous)
275
2
43
2
Hydrogen Fluoride
211
1.2
101
5
Another useful comparison is to compare the RMP and OCA data for a facility of interest
with RMPs and OCA data for facilities in industry sectors that had an accident history. Accident
history can be a useful indicator of the likelihood of release. Table 9 presents the 21 industry
sectors that reported the greatest number of accidents in the five-year accident history section of
the RMP. This data, along with the number of facilities reporting in Table 6, provides a measure
of likelihood that a facility in a particular sector will have an accident. For example, there is
approximately one reported accidental release in the five year RMP accident history for every 50
facilities reporting in the Agricultural Supply sector and less than one for every 10 facilities
reporting in the Sewage Treatment sector, while in the Petroleum Refineries sector there is more
than one per reporting facility.
Table 9 - RMP 5-Year Accident History by Industry Sector
Industry Sector Description - NAICS Code
Number of
Accidents
Percent of RMP
Accident History
Petroleum Refineries - 32411
190
10
Water Supply and Irrigation - 22131
116
6
Sewage Treatment - 22132
110
6
Basic Inorganic Chemical Manufacturing
(Not otherwise specified) - 325188
89
5
Basic Organic Chemical Manufacturing (Not otherwise specified) - 325199
89
5
Other Chemical and Allied Products - 42269
87
5
Agricultural supply - 42291
85
4
Alkalies and Chlorine Manufacturing - 325181
80
4
40
-------
Nitrogenous Fertilizer Manufacturing - 325311
68
4
Pou Itry Processing - 311615
67
4
Petrochemical Manufacturing - 32511
55
3
Pulp Mills- 32211
54
3
Refrigerated Warehousing - 49312
50
3
Animal (except Poultry) slaughtering - 311611
47
2
Natural Gas Liquid Extraction - 211112
34
2
Plastics Material and Resin Manufacturing - 325211
34
2
Frozen Fruit, Juice, and Vegetable Manufacturing - 311411
32
2
Meat Processed from Carcasses - 311612
31
2
Paper (except newspaper) mills - 322121
25
1
Industrial Gas Manufacturing - 32512
24
1
Other Basic Organic Chemical Manufacturing - 32519
24
1
CONCLUSION
The collection of enough information to adequately compare the hazards and risks
between facilities is a critical element in understanding the risk and eventual risk reduction at a
particular facility. Even though OCA data does not provide a complete picture of risk, a
comparison of the worst-case scenario and alternative scenario (in particular the distance to
endpoint for these scenarios) allows a greater understanding of the maximum extent (and a more
likely extent) of the hazard component of the risk associated with RMP reporting facilities.
Several types of comparisons are outlined in this chapter, and each one requires access to a
significant number of RMPs and OCA data in order to generate a complete comparison.
41
-------
CHAPTER 7
WHAT IS THE PUBLIC'S ACCESS TO OCA INFORMATION
UNDER CSISSFRRA?
This assessment has so far examined whether public access to TRI and other risk
information has resulted in risk reduction. It has also explored whether the nature of, and the
means of access to, the information has affected the public's use of the information. With respect
to OCA data in particular, the assessment has considered whether public access to that data
would likely result in chemical accident risk reduction.
EPA's charge is to assess the incentives created by public access to "off-site consequence
[OCA] information," which is related to, but not synonymous with, OCA data. CSISSFRRA
defines "[OCA] information" as "those portions of a [RMP], excluding the executive summary of
the plan, consisting of an evaluation of 1 or more worst-case release scenarios or alternative
release scenarios, and any electronic data base created by the Administrator from those portions"
(CAA section 112(r)(7)(H)(i)(III)). Notably, the definition refers to specified portions of RMPs
and any EPA database created from these portions, not to the data reported in those portions. It
also expressly excludes RMP executive summaries, which are required to include at least a brief
description of the information reported in the OCA portions of RMPs. In fact, most RMPs
include at least some OCA data in their executive summaries.
Relatedly, CSISSFRRA states that it "does not restrict the dissemination of [OCA]
information ... in any manner or form except in the form of a [RMP] or electronic data base
created by [EPA]" (emphasis added) (CAA section 112(r)(7)(H)(xii)(II)). Together these
provisions make clear that "OCA information" refers only to the specified forms of OCA data, not
the data itself. (The statute restricts these forms of OCA data because they are formatted in a
way that makes them fairly easy to compile into a large OCA database that could be posted on the
Internet.) Accordingly, EPA must ultimately assess the incentives that would be created by public
access to "OCA information," or OCA data in the restricted forms (sections 2 through 5 of RMPs
and any electronic database EPA creates from those sections).
CSISSFRRA provides the public with some means of access to OCA data (i.e., the data
reported in the OCA sections of RMPs) even before, or in addition to, the regulations that are to
govern distribution of "OCA information." As noted above, RMP executive summaries are
excluded from the definition of "OCA information" and thus from the statute's restrictions on
dissemination. Executive summaries are required to include a "brief description" of OCA data.
EPA's rule does not define a "brief description," leaving facilities to make reasonable decisions as
to what information to include. A random sampling of RMPs indicates that the amount of OCA
data included in executive summaries varies from facility to facility, with some facilities providing
nearly complete data and others providing little. However, most executive summaries provide at
least some OCA data. The summaries are already available on the Internet through multiple web
42
-------
sites.
Besides executive summaries, CSISSFRRA ensured early public access to at least
summaries of OCA data. It required virtually all covered facilities by February 1, 2000 to conduct
a public meeting or post a public notice summarizing the OCA sections of its RMP (CSISSFRRA
section 4). To date, the Federal Bureau of Investigation has received notification from
approximately 5000 facilities that they have complied with this requirement. In addition,
CSISSFRRA allows any facility to actually release the OCA sections of its RMP to the public
without restriction, and once a facility has done so, the information is no longer restricted (CAA
section 112(r)(7)(H)(v)(III)). To date, EPA has received notification from over 900 facilities that
they have released their OCA information without restriction.
CSISSFRRA also allows governmental officials to communicate OCA data to the public,
so long as they do so in a form that does not replicate the OCA sections of RMPs or EPA's OCA
database. CSISSFRRA guarantees governmental officials access to "OCA information" for their
"official use" (see CAA section 112(r)(7)(H)(iv) and (ii)(cc)-(ee)). Governmental officials,
referred to as "covered persons" by the statute, include officers and employees of federal, state or
local government or their agents or contractors, and officers and employees of state and local
emergency response officials or their agents or contractors (see CAA section 112(r)(7)(H)(i)(I)).
Emergency response officials include members of State Emergency Response Commissions
(SERCs) and Local Emergency Planning Committees (LEPCs) created under the Emergency
Planning and Community Right-to-Know Act (EPCRA). Members of these commissions and
committees can include members of the public, the media, and industry, as well as representatives
of emergency responders such as fire and police departments (EPCRA section 301(c)).
Emergency response officials, including fire fighters, are "covered persons" whether or not they
are paid for their services.14
While CSISSFRRA guarantees covered persons access to OCA information, it prohibits
them from disclosing the information to the public except as authorized by the statute or the
regulations issued under it (section 112(r)(7)(H)(v)). It also prohibits them from disclosing "any
statewide or national ranking of identified stationary sources derived from" OCA information.
Any covered person who violates the prohibition is subject to criminal penalties of up to
$1,000,000 for violations committed in any one year.
Notwithstanding these prohibitions, CSISSFRRA "does not restrict the dissemination of
[OCA] information by any covered person in any manner or form except in the form of a [RMP]
or an electronic data base created by [EPA] from off-site consequence analysis information," as
noted above. Thus, CSISSFRRA prohibits disclosure of RMP sections 2 through 5, or OCA data
conveyed in the "form" of those sections, and prohibits disclosure of EPA's OCA database. But
it does not prohibit disclosure of OCA data when the data is disclosed in a form different than that
14 "Covered persons" also include "qualified researchers" under CAA section
112(r)(7)(H)(vii).
43
-------
of RMP sections 2 though 5 or EPA's OCA database. Congress so limited the scope of the
prohibition so that governmental officials could communicate to the public about the potential off-
site consequences of chemical accidents, but in a way that does not lend itself to Internet
dissemination. (See, e.g., 145 Cong. Rec. H6083 (July 21, 1999) (statement of Rep. Dingell).)
Covered persons may consequently convey to the public the data in RMP sections 2 through 5
and EPA's OCA database, so long as they do not hand out copies of, or otherwise replicate, the
restricted RMP sections or provide access to EPA's database.
Finally, CS1SSFRRA guarantees public access to OCA information itself (i.e., the OCA
sections of RMPs and EPA's database created from those sections) in several specified ways apart
from the regulations governing distribution of OCA information. First, it requires EPA to make
OCA information available to the public without information concerning the identity and location
of facilities reporting the information (CAA section 112(r)(7)(H)(iv)). That information is
currently available upon request. Second, it requires EPA, in consultation with DOJ and other
agencies, to establish a "read-only information technology system" that "provides for the
availability to the public of [OCA] information by means of a central data base under the control
of the Federal Government that contains information that users may read, but that provides no
means by which an electronic or mechanical copy of the information may be made" (CAA section
112(r)(7)(H)(viii)). EPA is working with other federal agencies to develop this read-only system.
Third, CSISSFRRA requires EPA, in consultation with DOJ, to make OCA information available
to "qualified researchers" by means of a system that does not allow researchers who receive the
information to disseminate it (CAA section 112(r)(7)(H)(vii)). EPA expects to initiate this system
soon.
Against this backdrop of guaranteed but limited access, EPA must evaluate "the incentives
created by public disclosure of [OCA] information for reduction in the risk of accidental
releasesf.]" EPA has so far assessed the incentive that public access to OCA data would create
for chemical accident risk reduction. EPA must now consider the importance of public access to
OCA information - the restricted forms of OCA data - to the creation of that incentive, in light of
the public access CSISSFRRA already provides.
Guaranteed access to OCA information has several important advantages over the access
to OCA data that CSISSFRRA otherwise provides. To begin with, OCA information provides
full OCA data for covered facilities. RMP executive summaries, by contrast, communicate only
as much OCA data as facilities choose to include. While a minority of facilities included nearly
complete OCA data, others provided little. A person interested in a particular facility may or may
not find the OCA data of interest to that person in the facility's executive summary.
In the public meeting or notice required by CSISSFRRA, facilities were required to
summarize OCA data, not provide OCA data itself. Further, facilities had discretion regarding
how to summarize OCA data.
Under CSISSFRRA, governmental officials are another potential source of OCA data, but
whether they communicate it is left to their discretion. EPA is so far unaware of any
44
-------
governmental officials communicating OCA data to the public. Indeed, few governmental
officials have even requested access to OCA information to date. EPA has learned from several
members of LEPCs and other state and local officials that they are reluctant to obtain OCA
information out of concern for the criminal penalties associated with unauthorized distribution of
it. CSISSFRRA only punishes willful violation of its disclosure restrictions, but these officials
have nonetheless expressed concern that inadvertent disclosure might result in criminal fines. (32)
In any event, if governmental officials do not have OCA information themselves, they cannot
communicate OCA data to the public. More fundamentally, providing governmental officials with
discretion to convey OCA data leaves up to them whether and to what extent they provide access
to the data. And even if government officials choose to communicate OCA data, the public
cannot be sure they were communicating it accurately without access to OCA information itself.
Facilities may also disclose their OCA data to the public, but they, too, have discretion
regarding whether to do so. Questions of accuracy may also arise to the extent facilities release
something other than the OCA sections of their RMPs. To date, only a small percentage of
facilities have released the OCA sections of their RMPs without restriction.
Given these limitations on existing public access to OCA data, guaranteed access to OCA
information would assure members of the public that they can receive complete and accurate
OCA data for the facilities of interest to them.
Guaranteed access to OCA information would also ensure that the public could gain
access when they needed it. Governmental officials and facilities may or may not choose to
communicate OCA data at any given time. CSISSFRRA required facilities to provide only one
public meeting or notice by a date now past. Initial indications are that relatively few members of
the public attended the meetings. For persons who missed the meeting or notice or moved to the
area afterwards, there is no subsequent opportunity to obtain at least a summary of OCA data
from the facility itself.
OCA information, by virtue of its format, has the further advantage of putting OCA data
into context within an RMP. When presented as part of the RMP, OCA data can be reviewed
together with information about a facility's accident history and prevention program. When
considered along with RMP information, OCA data provides greater insight into the risk a facility
poses and the steps the facility is taking to manage or reduce that risk.
In addition to providing some access to OCA data, CSISSFRRA guarantees that OCA
information itself will be available to the public in several prescribed ways independent of the
regulations. However, these avenues of access have significant drawbacks. OCA information
without facility identification or location information is of limited utility. The most obvious
limitation is that such a database does not allow a member of the public to know which facility's
data he or she is viewing. A more subtle limitation has to do with the fact that RMP reporting is
general in nature. RMPs do not explain how a facility is using prevention measures to reduce
risk, only that certain types of measures are being used. For example, a facility need only report
that it uses shut-off valves and a dike to reduce the volume and dispersion of an accidental
45
-------
release, but it does not have to provide the specifications or locations of those measures. For a
member of the public to fully understand a facility's prevention program, he or she must contact
the facility (or ask a governmental official or other representative to do so) to gain more specific
information. The limited usefulness of OCA information without facility identification data is
perhaps best indicated by the fact that, to date, no one has requested the information.
The read-only OCA database required by CSISSFRRA is also likely to have drawbacks.
It is unclear at this time how such a read-only system will operate and how many outlets for the
system will be available. Government resources for the read-only technology system are limited.
Finally, the required system for access to OCA information for qualified researchers is, by
definition, not available to the general public. While the system may result in risk information
being developed and communicated to the public by researchers, it does not provide the public
with access to OCA information itself and it precludes researchers from providing that access.
In summary, CSISSFRRA already makes OCA data, and even OCA information (the
restricted forms of OCA data), available to some extent. EPA has thus evaluated the incremental
benefit of guaranteeing public access to OCA information. While the public has access to at least
some OCA data through RMP executive summaries and may obtain additional OCA data as a
result of other, discretionary actions by governmental officials or facilities, the extent to which the
public receives OCA data through these avenues ultimately depends on decisions made by others.
CSISSFRRA's provisions for public access to OCA information without facility identification
information and in read-only database are limited in nature and thus effect. Without guaranteed
access to OCA information as such, a member of the public may or may not obtain the OCA data
for the facilities of interest to her or him. And without that data, she or he would have no basis -
or incentive - to initiate actions directed at reducing the risk of chemical accidents. Furthermore,
a lack of data on the part of the public would mean that facilities, including those that lag behind
their peers, would have fewer incentives to reduce the risk of chemical releases.
46
-------
CHAPTER 8
FINDINGS
The President delegated to EPA the responsibility for assessing the incentives for
reduction in chemical accidents created by public disclosure of OCA information. This
assessment has produced a number of findings that address the relevant issues.
Chemical accidents impose substantial costs on the American public and on industry.
Although catastrophic chemical accidents that kill many people at once are fortunately
relatively rare, chemicals accidents continue to cause deaths, injuries, property damage,
disruption of lives, and business losses. These accidents continue to occur at facilities that
are subject to accident prevention regulations. Almost 80% of the serious accidents
reported in the RMP five-year accident history occurred at facilities subject to the OSHA
PSM standard. Although it is likely that the PSM standard has prevented some releases,
there is clearly a need for additional efforts to improve safety at facilities handling highly
hazardous chemicals.
• Public information does result in the public acting on the information, and those actions
have very likely led to risk reduction. Public use of the TRI data has included reports and
campaigns from public interest groups at the local, state, and national level, press
coverage, and state legislation. It is not possible to quantify the level of risk reduction
produced by such actions with any certainty. TRI emissions have decreased by 43% since
1988, although other factors have produced some of that reduction. Negative press
coverage directed at certain facilities appears to have led these facilities to achieve
reductions in their TRI emissions.
Over the last 10 years, reportable releases of hazardous substances from facilities in four
states reviewed have declined by 68 percent while reports of oil spills and transportation
releases have showed no consistent trend. All the reasons for the decline in reportable
releases are not known, but one likely factor is that hazardous substances are the subject
of public scrutiny to a far greater degree than oil and transportation, leading facilities to
improve their management of these chemicals to reduce risks.
• The type of information provided to the public affects its use. Scorecard shows that the
public is at least an order of magnitude more likely to access interpreted data than it is to
seek raw data. This is because interpreted data, such as OCA data, are put into context
and made more understandable by the public.
• Ease of access is important. The greater the effort needed to obtain data, the less likely
members of the public will obtain it. EPCRA section 312 data that are available locally
have rarely been used (about 3,500 requests per year). TRI and related data on the
47
-------
Internet, even without interpretation, draw 240,000 searches a year at one Internet site.
With interpretation and easy access, Internet users view TRI and other data over seven
million times per year at another website.
• Current public access to OCA data is uneven, and consequently is ineffective relative to
consistent, broad access. Executive summaries vary widely in the amount of OCA data
reported. Public officials have been reluctant to even access the data because of liability
concerns; there has thus been little chance for the public to learn OCA data from them.
OCA could be derived, though not accurately, from the raw data, but the technical
challenges are likely to prevent the public from doing so.
• Actual chemical releases are different from the releases evaluated in the OCA information,
but, because facilities generate OCA information according to consistent assumptions, the
public and others can understand and compare the relative potential hazards and risks
present at a facility.
As noted at the outset of this assessment, chemical accidents continue to claim lives,
health and property. The evidence and analysis set forth in the assessment demonstrate that
providing the public with reasonably convenient access to OCA information would likely result in
significant reductions in the risk of chemical accidents over and above the risk reduction that the
RMP rule on its own will accomplish. Conversely, to the extent the public is not given workable
access to OCA information, the risk of chemical accidents is not likely to be reduced as much as it
world if such access were provided. And the loss of that increment in risk reduction would mean
the loss of lives, health and property that could have been saved.
48
-------
REFERENCES
1. Loss Prevention in the Process Industries, Second Edition, F.P. Lees, Butterworth-
Heinemann Publishing, 1996, App. 5/9.
2. Loss Prevention in the Process Industries, App. 4/2.
3. Clean Air Act, 42 U.S.C. section 7412 (r).
4. U.S. Chemical Safety and Hazard Investigation Board, "Chemical Safety Board To
Conduct Comprehensive Investigation of Pennsylvania Chemical Plant Explosion," March
3, 1999.
5. Economic Analysis in Support of Final Rule on Risk Management Program Regulations
for Chemical Accident Release Prevention, As Required by Section 112(r) of the Clean
Air Act, May 21, 1996.
6. Adams, W., S. Burns, and P. Hardwerk. National LEPC Survey, October 1994. George
Washington University, App. A, Table 25.
7. U.S. Government Accounting Office, "Environmental Information: Agencywide Policies
and Procedures Are Needed for EPA's Information Dissemination," GAO/RCED-98-245,
Sept. 1998.
8. Lynn, F. M. and J. Kartez. "Environmental Democracy in Action: the toxics release
inventory," Environmental Management, Vol. 18, No. 4, pp. 511-521, 1994.
9. 1997 Toxics Release Inventory, US EPA, EPA 745-R-99-003
10. World Bank, "Greening Industry: New Roles for Communities, Markets, and
Governments," Oxford University Press, 1999, pp 64-74.
11. "With Toxic Risk, Plans Vary," Washington Post, Oct. 10, 1999, p CI, C10-C11.
12. "Plant Warnings Go Unheeded," Washington Post, Nov. 5, 1999, p Al.
13. "Blue Plains Details Safety Plans," Washington Post, March 3, 2000, p B3.
14. Robert Barrish, Delaware Department of Natural Resources and Environmental Control,
Division of Air & Waste Management, personal communication, with David Wiley, EPA,
Dec. 20, 1999.
15. Mark Zusy, Nevada Department of Environmental Protection, Bureau of Waste
49
-------
Management; personal communication with Craig Matthiessen, EPA, January 2000.
16. National Institute of Chemical Studies, NICS News, Vol. 8, Fall 1999, p 6.
17. National Highway Traffic Safety Administration, "Consumers need more reliable
automobile fuel economy data," Report No. CED-81-133, July 29, 1981
18. Henderson, David., "The Economics of Fuel Economy Standards," RGO, Vol. 9, No. 1,
Jan./Feb. 1985, pp 45-48.
19. Neuhouser, Kxistal, and Patterson. "Use of food nutrition labels is associated with lower
fat intake," Journal of the American Dietetic Association, 99, 1, 45(6), Jan., 1999.
20. Correspondence with Dr. William S. Pease, Director, Internet Projects, Environmental
Defense Fund, 11/22/99 and 2/22/00.
21. Wolf, Sidney. "Fear and loathing about the public right to know: the surprising success of
the Emergency Planning and Community Right-to-Know Act." Journal of Land Use and
Environmental Law, V. 11, No. 2, Spring 1996.
22. Adams, W., S. Burns, and P. Hardwerk.
23. Reginald Baldini, NJ Department of Environmental Protection, personal communication,
with David Wiley, EPA, Dec. 8, 1999.
24. Robert Barrish, Dec. 20, 1999.
25. Mark Zusy, January 2000.
26. Technical Guidance for Hazards Analysis - Emergency Planning for Extremely
Hazardous Substances; US EPA, US FEMA, US DOT; December 1987.
27. Planning for the Worst, Maps; The Augusta Chronicle Online, 10/9/97; Augusta, Georgia;
www.augustachronicle.com.
28. Loss Prevention in the Process Industries, Second Edition, F.P. Lees, Butterworth-
Heinemann Publishing, 1996, pp 2/2-3.
29. Safety in the Chemical Industry, Lessons from Major Disasters, E.A. Stallworthy and
O.P. Kharbanda, G.P. Publishing, Columbia, MD, 1988, pp 99-100.
30. Guidelines for Evaluating the Characteristics of Vapor Cloud Explosions, Flash Fires, and
BLEVEs, CCPS/AIChE, 1994, pp 70-75.
50
-------
31. Safety in the Chemical Industry, pp 99-100.
32. Timothy Gablehouse, Chair, Jefferson County Colorado LEPC and LEPC representative
to the Federal Accident Prevention Advisory Subcommittee, personal communication with
David Wiley, EPA, Dec. 2, 1999.
33. Graham, Mary; Regulation by Shaming-, The Atlantic Monthly, April 2000!
34. Right-to-Know Planning Guide, "EPA Notices Change in Users of Toxic Release
Inventory." Bureau of National Affairs. (February 1990): p. 4.
35. Tryens, Jeffrey, Richard Schrader, and Paul Orum. Making the Difference: Using the
Right-to-Know in the Fight Against Toxics. Washington, DC: Center For Policy
Alternatives and Working Group on Community Right-to-Know. (undated): p. 2.
36. ICF, Inc., Memorandum: "Transmittal of Research Materials on Negative Press," April
18,2000.
51
-------
This page intentionally left blank.
52
-------
APPENDIX A
BACKGROUND
OF
EPA'S RISK MANAGEMENT PROGRAM
AND
OCA INFORMATION
A-l
-------
APPENDIX A
BACKGROUND OF EPA'S RISK MANAGEMENT PROGRAM AND OCA
INFORMATION
This Appendix provides background information on the accident prevention provisions of
the Clean Air Act, describes EPA's Risk Management Program and the information contained in a
Risk Management Plan (RMP), and discusses what Off-site Consequence Analyses are, including
what information is included in the OCA portion of a risk management plan, and what is not
included.
1. What Role Do the Clean Air Act Amendments of 1990 Play in Accident Prevention?
The CAA Amendments of 1990 authorized regulations and programs to prevent accidental
chemical releases and to minimize the consequences of such accidental releases when they occur.
The sections added to the CAA for this purpose are:
Section 112(r)(l) - establishes a general duty on facilities handling any extremely
hazardous substance to identify hazards which may result from accidental releases using
appropriate hazard assessment techniques, to design and maintain a safe facility, and to
minimize the consequences of accidental releases which do occur.
• Sections 112(r)(3), (4) and (5) - require EPA to establish a list of at least 100 substances
that pose the greatest risk of causing death, injury, or serious adverse effects to human
health or the environment from accidental releases along with a threshold amount for each
substance.
• Section 112(r)(6) - establishes a Chemical Safety and Hazard Investigation Board to
investigate accidental releases and advise EPA and the Department of Labor, Occupational
Safety and Health Administration (OSHA) on the efficacy of their regulatory programs.
• Section 112(r)(7) - directs EPA to issue reasonable regulations and appropriate guidance
to prevent and detect accidental releases and to require facilities1 with more than a
threshold amount of a substance listed under sections 112(r)(3)-(5) to develop and
implement risk management plans (RMPs). This section specifies that RMPs include an
evaluation of the off-site consequences of worst-case releases; the RMPs, including the
consequence evaluations, must be submitted to the Chemical Safety Board and state and
local officials and be made available to the public.
' The CAA and the regulations use the term "stationary source" rather than "facility,"
which is used in EPCRA. These terms are synonymous; this report generally uses the term
"facility."
-------
• CAA Amendment section 304 - requires the Department of Labor to promulgate
regulations under the Occupational Safety and Health Act to establish a chemical process
safety management (PSM) standard designed to protect employees from hazards
associated with accidental releases of highly hazardous chemicals in the workplace. This
standard was issued by OSHA in 1992. It requires facilities having more than a threshold
quantity of certain highly hazardous chemicals listed by OSHA to implement and
document an accident prevention program. For example, facilities subject to OSHA PSM
must prepare and use written operating and maintenance procedures for hazardous
chemical processes, must conduct a systematic analysis of process hazards and ensure that
all hazards are controlled, and must conduct training for workers who operate hazardous
chemical processes. The elements of the OSHA PSM accident prevention program also
serve as the core accident prevention elements for EPA's Risk Management Program.
2. What are the Requirements of EPA's Risk Management Program Regulation?
EPA addressed the chemical accident prevention and detection requirements in
112(r)(7)(B)(i) along with the risk management plan requirements in (B)(ii) in one regulatory
effort. In so doing, EPA required that facilities handling more than the threshold quantity of a
substance listed under sections 112(r)(3)-(5) must develop and implement a Risk Management
Program that evaluates the hazards present at the facility and establishes a chemical accident
prevention and emergency response programs. The most important feature of the Risk
Management Program is the chemical accident prevention program which is based on the
elements of chemical process safety management. The most effective way to prevent catastrophic
chemical accidents is through process safety management. The elements of process safety
management were derived by industry, trade associations, and professional societies. Chemical
process safety management calls for a systematic and rigorous evaluation of the chemical and
process hazards present in the facility and it brings together all the necessary elements for the safe
operation of that process, day-after-day, under one management system. Through process safety
management, facilities design, install, maintain, and operate the equipment necessary to prevent
and detect accidental releases.
The elements of process safety management were adopted by OSHA into its PSM
standard promulgated in 1992. EPA chose to adopt and build on OSHA's requirements and the
industry approach for its chemical accident prevention program because it is the most effective
way to prevent accidents, most chemical accident prevention actions taken to protect workers will
also protect the public, and separate requirements would be more burdensome, duplicative and
less proven than process safety management.
While EPA's Risk Management Program borrows much from the OSHA PSM standard,
its requirements extend beyond the OSHA standard. In addition to the PSM accident prevention
requirements, the EPA Risk Management Program requires facilities to: analyze the off-site
consequences of accidental releases, provide information about accidents that the facility has
suffered during the previous five years, and provide information about their accident prevention
and emergency response programs in a publicly-available risk management plan (RMP). These
A-3
-------
additional elements were intended, in large part, to use "the power of public scrutiny to promote
voluntary hazard reduction, often achieving far more benefits than what regulatory programs
could achieve on their own" (Legislative history of the Chemical Safety Information, Site
Security, and Fuels Regulatory Relief Act, Senate Report 106-70, June 9, 1999, page 12). By
requiring companies to analyze the potential off-site consequences of accidental releases and
including this analysis in the RMP, Congress and EPA predicted that the "right-to-know effect,"
when applied to the chemical industry by means of publicly available RMPs, would contribute to
an atmosphere in which industry, through non-regulatory means, sees incentives to take all
reasonable steps to reduce chemical risks.
In the Preamble to the final RMP rule published in the Federal Register on June 20, 1996,
EPA discussed the two underpinnings of EPA's approach to the risk management program. First,
EPA stated that with this rule "EPA continues the philosophy that the Agency embraced in
implementing the Emergency Planning and Community Right-to-K.now Act of 1986. Specifically,
EPA recognizes that regulatory requirements by themselves, will not guarantee safety. Instead,
EPA believes that information about hazards in a community can and should lead public officials
and the general public to work with industry to prevent accidents. EPA intends that officials and
the public use this information to understand the chemical hazards in the community and then
engage in a dialogue with industry to reduce risk."
Secondly, the Agency stated that the rule "builds upon existing programs and standards.
For example, EPA coordinated with OSHA and the Department of Transportation (DOT) in
developing this regulation. To the extent possible, covered sources will not face inconsistent
requirements under these agencies' rules." This approach was supported by public comments
which stated that sound process safety management systems ideally address chemical accident
prevention in a way that protects workers, the public and the environment.
In the Federal Register on January 31,1994, and in subsequent amendments, EPA
published a list of regulated substances under the risk management program. This list currently
contains 77 substances listed because of their volatility and acute toxicity, and 63 substances listed
because of their high flammability. Together, these substances represent the initial list of
regulated substances that are known to cause, or may be reasonably anticipated to cause, death,
injury, or serious adverse effects to human health or the environment if accidently released.
In the same notice, EPA published threshold quantities for each of the regulated
substances. The approach taken to set the threshold quantities focuses on the quantity of a
substance that might be released in a single accident, and that could be reasonably anticipated to
cause severe health effects as a result of an accidental release. Threshold quantities for toxic
substances range between 500 to 20,000 pounds. The threshold quantity for all flammable
substances is 10,000 pounds.
Any facility having more than a threshold quantity of a regulated substance in a process is
required to develop and implement a risk management program. However, EPA scaled the risk
management program requirements according to the relative risk posed by a facility to the
A-4
-------
surrounding community. Each process at a facility that has had no serious accidents in the past
five years and that can demonstrate that the worst-case accident scenario does not affect any
public receptors is eligible for "Program 1" requirements, which are the least stringent. Processes
at facilities not eligible for Program 1 that are already subject to OSHA PSM or belong to
industry sectors having a significant history of accidents are required to meet "Program 3"
requirements, which are the most extensive. Finally, processes that are not eligible for Program 1
or subject to Program 3 are required to meet "Program 2" requirements. Program 2 requirements
are similar to Program 3 except that they contain streamlined accident prevention program
measures (the OCA, accident history, and emergency response program requirements are the
same in Programs 2 and 3).
The risk management program must include:
A management system - Owners or operators of facilities with Program 2 or Program 3
processes must develop a system to oversee the implementation of the risk management
program elements, and assign a qualified person or position that has the overall
responsibility for development, implementation, and integration of the program elements.
Program 1 processes are not subject to the management system requirements.
An accident prevention program - For Program 3 processes, this is virtually identical to
the requirements under OSHA PSM, including measures such as written operating and
maintenance procedures, process hazard analysis, a mechanical integrity program, incident
investigations, compliance audits, and others. For Program 2 processes, the accident
prevention program contains a streamlined subset of the full PSM requirements. Program
1 processes are not subject to any additional prevention program requirements.
Prevention program requirements for Program 2 and 3 processes are "performance-
based," as opposed to "command-and-control," because chemical facilities and processes
are unique and therefore accident prevention programs must necessarily be tailored to
each facility.
An emergency response program - requires facilities to have procedures in place to notify
emergency response officials in the event of an accident and to coordinate with local
response agencies and community response plans. If a facility responds to its own
emergencies, facilities must have a written plan containing procedures for informing
emergency response agencies about emergencies, documentation of proper first-aid and
emergency medical treatment, procedures for emergency response to an accidental release,
including use of equipment, and training for employees who will respond to releases.
• A hazard assessment program that consists of a five-year accident history and an analysis
of the consequences of worst-case and other accidental releases:
1. The five-year accident history includes a description of prior accidental releases which
meet certain severity triggers, such as deaths, injuries, significant property damage,
evacuations, environmental damage, or sheltering in place.
A-5
-------
2. The off-site consequence analysis (OCA) includes an analysis of the potential
consequences of hypothetical worst-case and alternative release scenarios. Worst-case
scenarios assume the release of the greatest amount of a regulated substance held in a
single vessel or pipe under specified ambient and process conditions, taking into account
administrative controls that limit the maximum quantity, and accounting for the effects of
passive mitigation features if present. Alternative release scenarios assume a release that
is more likely to occur than the worst-case, using release parameters chosen by the facility
owner as appropriate for the scenario, and may account for both passive and active
mitigation features.
As required by Congress in section 112(r), all facilities subject to the RMP regulation,
regardless of Program level, must submit an RMP that documents elements of their risk
management program. RMP contents are described in detail in the next subsection.
Facilities subject to the program were required to submit an initial RMP by June 21, 1999,
and must submit an update at least every five years. Facilities must update their RMP sooner
under certain circumstances (e.g., major change involving a regulated substance or process at the
facility). Approximately 15,000 facilities have submitted an RMP to date.
3. What information is reported in a Risk Management Plan?
The risk management plan (RMP) is intended to provide information that can be used by
others to judge the risk that a facility poses to the surrounding community and to understand the
steps taken by that facility to manage that risk. (A fictitious sample RMP is shown in section
15 below.) The executive summary is an overall text description of a facility's risk management
program, including, in general terms, the potential off-site consequences of the accidental releases
from the facility. The rest of the data in the RMP generally consists of yes/no, check-off box, and
numerical answers to standard questions. There are additional areas where facilities may include
text explanations for various entries, but (with the exception of the executive summary) these are
optional. The advantage of this format is that it allows data to be easily submitted, compiled and
managed in electronic form. However, also as a result of this format, information submitted in
RMPs is usually not extremely detailed. For example, a facility would indicate, by checking
various choices in a list, what types of mitigation measures it uses in a process, but unless the
facility chooses to add an optional explanation, the reader can not discern details such as precise
locations, methods of operation, or design features of those devices. Facilities are, however,
required to maintain on-site documentation which supports the information contained in the RMP
and implementation of the overall risk management program.
RMPs contain the following sections. The presence of some sections and the total number
of pages for RMPs vary depending on the number and type of processes and chemicals present at
a facility.
• Section 1: Registration information (e.g., facility name, address, process chemicals, etc.)
and an executive summary which provides a brief description of the accidental release
A-6
-------
prevention and emergency response policies at the source, worst-case and alternative case
scenarios, five-year accident history, and planned changes to improve safety.
Sections 2-5: Evaluation methodology and data for the off-site consequence analyses of
worst and alternative release scenarios. These sections provide data on the possible
consequences of the scenarios, as well as the assumptions and models used to obtain this
data.
Section 6: Five-year accident history data. For each past accident, the facility provides the
date of the event, chemical(s) released, source of release, on-site and off-site impacts,
initiating event, and factors contributing to the release.
Section 7: Contains a description and data for the processes subject to prevention
Program 3. Besides an optional narrative on the prevention program, facilities are
required to provide such information as the date of the last process hazards analysis, the
major hazards identified by that analysis, process controls used to address these hazards,
and information on maintenance, training, compliance audits, and incident investigations.
Section 8: Like Section 7, this section contains a description and data for processes
subject to prevention Program 2 .
Section 9: Contains data on the facility's emergency response program and plan.
4. How is RMP Information Managed? How Did EPA Intend for the Public to Obtain
Access?
EPA learned from its experience with the EPCRA program (see Chapters 3 and 4) that
electronic submission of data has several benefits over the submission of paper forms. First,
electronic submissions reduce the burden on regulated and receiving entities. Second, the
Agency noted that local agencies often lack the resources needed to make use of the information
reported to them by industry under federal programs, so the Agency wanted to limit the
information management burden on local entities so they could focus on the chemical safety issues
raised by the data. Third, EPA learned that electronic submissions would benefit affected
communities and the general public. The Agency believed this type of submission would promote
consistency and uniformity to enable communities and the general public to better understand the
data. When the agency proposed this method of data management, most of the public comments
supported the proposal to submit RMPs in electronic form to a central location, and EPA adopted
this approach when it promulgated the RMP rule (61 Fed. Reg. at 31673, 31694-95).
To help implement EPA's decision to develop a centralized management system for
RMPs, EPA convened a subcommittee under the Federal Advisory Committee Act, which
included representatives from state and local governments (including LEPCs), academia, industry
and public interest groups. One of the major issues the Accident Prevention subcommittee
A-7
-------
considered was how RMPs should be submitted and how RMP information could be managed.
The subcommittee unanimously agreed with EPA that RMPs should be submitted electronically,
that EPA should compile the RMPs into a central electronic database, and that EPA should make
that database available to Chemical Safety Board and state and local officials.
In a Supplemental Notice of Proposed Rulemaking, EPA requested comments on how
public participation in the risk management program process could be encouraged. EPA's
preferred approach was to encourage the public and sources to use existing groups, primarily
LEPCs, as a conduit for communication between the source and the public throughout the RMP
development process. While a substantial number of commenters supported this approach, many
opposed it because some LEPCs are not functional and because LEPCs were not seen as an
appropriate substitute for public participation. In the final rule, EPA did not adopt any specific
public participation requirements, but rather decided at the time to make all of the RMP
information immediately available to the public. EPA believed that by doing this, people would be
able to compare facilities in their community with similar facilities in other areas, and thereby gain
a better understanding of local industries in order to carry on a more informed dialogue with
facilities about their hazards and accident prevention practices.
The Accident Prevention subcommittee also unanimously agreed that the electronic RMP
database, except for the OCA information, should be made available to the public over the
Internet. Most subcommittee members believed that the OCA portion of the database should also
be available to the public on the Internet, but one member expressed concern that placing the
OCA information on the Internet could provide a targeting tool for criminals and terrorists. Law
enforcement agencies shared this concern, which evolved into Public Law 106-40. The public
currently has unrestricted access via the Internet to all sections of the RMP database except the
OCA information. EPA's database of OCA information is currently withheld from the public in
accordance with Public Law 106-40.
Finally, as stated previously, more than any other information reported in an RMP, OCA
information provides an easily understood means of evaluating the hazards a particular facility
poses to its surrounding community and how its hazards compare to those of similar facilities.
The data tell the public how far a worst-case or alternative scenario release from a particular
facility could travel, roughly how many people could potentially be affected, and what types of
"public receptors" (e.g., homes, hospitals, schools, businesses, parks) could be in the path of a
release. In short, the data allow members of the public to determine whether they could be
harmed by a chemical release from a particular facility.
5. What is OCA information?
Off-site Consequence Analysis (OCA) information is the portion of an RMP that contains
analyses of the possible consequences to a surrounding community of hypothetical chemical
accident scenarios. While these scenarios are not meant to predict exactly what will occur in an
actual event, they are an attempt to determine what could occur under certain conditions, and to
present this information in an easily understood fashion.
A-8
-------
The OCA is a consequence assessment, not a risk assessment. Risk assessments typically
measure risk as the product of the hazard multiplied by its probability (or frequency). For
example, the risk of being struck and killed by lightning is a function of the hazard (lightning's
potential to cause death) and the likelihood that if you are outdoors during a thunderstorm, that
you will be struck. Accidental release risk considers the hazard of the chemical accidentally
released (e.g., it's toxicity), the consequences of that release (how much will you be exposed to it)
and the likelihood that the chemical will be accidentally released. The OCA provides a rough
estimate of only the hazards and consequences of an accidental release, without evaluating the
likelihood or probability that such an accident will occur. Since the likelihood of an accidental
release varies considerably from facility to facility, the RMP requirements, in effect, assume that
the likelihood of a worst-case release or alternative case release is equal across all facilities. This
simplifies the analysis and allows comparability of hazards and consequences across RMP
facilities. The consequences are expressed in terms of the potentially affected population, as well
as the types of buildings, parks, and other public and environmental areas that could be seriously
affected by a release.
The OCA requirements of EPA's Risk Management Program uniquely distinguish it from
any other federal regulatory program. Each facility subject to the RMP regulations must conduct
an OCA, report the results of the analysis in their RMP submission, and in most cases, discuss the
analysis with the local public. In an OCA, all sources are required to develop at least one worst
case scenario. Additionally, most sources (all except Program 1 sources) must also develop at
least one alternative release scenario.
6. What is a worst-case scenario?
Worst case scenarios are highly unlikely accident scenarios that are intended to serve as a
measure of the maximum hazard that a chemical facility could pose to the surrounding area. They
assume that the facility accidentally releases the entire contents of its largest tank or pipe of toxic
or flammable material into the environment under very stable atmospheric conditions. The
scenario also assumes that any active release mitigation systems (i.e., systems that require human,
mechanical, or energy input to function) fail to operate, but that passive systems do operate.
Stable atmospheric conditions (i.e., low wind speed and high atmospheric stability) are assumed
because they are most conducive to lengthening the distance that a highly concentrated toxic gas
cloud travels as it moves outward from its source; breezy, unstable conditions cause the cloud to
disperse in a relatively short distance. For flammable worst case scenarios, the analysis assumes
that a vapor cloud explosion occurs after the release. A vapor cloud explosion is the type of
flammable gas accident that could generally affect the greatest geographic area.
7. What is an alternative release scenario?
Alternative release scenarios are more likely, and generally less severe, than worst case
scenarios. These scenarios are intended to be a more realistic estimate of the consequences to the
surrounding community of a chemical accident at a given facility. Many facilities and local
emergency planners use them to prepare emergency response plans. Alternative scenarios are
A-9
-------
selected by the facility based on the facility's accident history, hazards analysis, or the experience
and judgement of the owners or operators of the facility. Alternative scenarios incorporate more
realistic assumptions in their analysis than worst-case scenarios. Alternative release scenarios
would generally assume that a smaller accidental release occurs under typical atmospheric
conditions and that passive and active release mitigation systems operate properly. For example,
an alternative scenario might be a 10-minute leak from a split in a transfer hose on a breezy, partly
cloudy day.
8. How are OCA scenarios developed and how accurate are they?
Both worst-case and alternative release scenarios are hypothetical estimates. They are
analyzed using mathematical vapor cloud dispersion models (for toxic scenarios) or fire/explosion
models (for flammable scenarios) to predict either the extent that a toxic gas cloud would spread
or the extent of blast or radiant heat effects from an explosion or fire of highly flammable material.
Each scenario includes an estimated distance outward from the source that may be subject to
concentrations, over pressures, or high temperatures of a toxic substance release or flammable
chemical explosion or fire that could cause irreversible acute health effects or death to human
populations within that range. The analyses are based on estimates of the quantity of a chemical
released, the rate of release, airborne dispersion and the airborne concentrations (for toxics) or
blast effects (for flammable substances) that could cause at least irreversible health effects.
Many valid methods are available to conduct the OCA. These methods usually involve
using either computer programs or lookup tables in which the analyst enters various parameter
values (e.g., atmospheric conditions, terrain roughness, etc.), and the computer program or
lookup table (also based on a computer simulation) provides the scenario endpoint, or
consequence distance. For worst-case scenarios, most input assumptions are specified by
regulation in order to provide some basis for comparison among similar sources. However, other
assumptions are facility-dependent and are therefore selected by the facility analyst (for this
reasons and others explained below, the OCA results from two similar sources may differ greatly).
In alternative scenarios, all parameters are selected by the facility analyst.
The results of OCA scenarios are rough estimates that are generally conservative (i.e.,
likely to over-predict actual consequences). The uncertainty in the estimates arises from several
factors, including the fact that actual atmospheric conditions at the time of an accidental release
will be unknown, process conditions may change from those selected for analysis, and because the
science of modeling large gas releases over long distances is highly complex and still evolving.
Results will usually over-predict actual consequences for two reasons. First, the rule-specified
assumptions for worst-case scenarios represent conditions that are most conducive to causing
severe off-site effects. However, these conditions (i.e., high atmospheric stability and low wind
speed for toxics, 10% explosive yield factor for flammable scenarios), are extremely rare. Any
change from these conditions during an actual release will generally reduce the consequence
distance. The second reason that OCA scenarios usually over predict consequences is because of
the uncertainty in our understanding of some of the physical mechanisms involved in atmospheric
dispersion and blast propagation. This has generally led scientists to err on the side of caution
A-10
-------
when developing models.
9. Exactly what data appears in the OCA information of an RMP?
Facilities are required to analyze and report in the RMP one worst-case release scenario
for each Program 1 process, one worst-case scenario for a toxic substance, one worst-case
scenario for a flammable substance, and additional worst-case scenarios if a worst-case release
scenario can affect a different public receptor then the other worst-case scenarios listed above.
Sources are required to identify and analyze at least one alternative release scenario for each
regulated toxic substance held in a covered process(es) and at least one alternative release
scenario to represent all flammable substances held in covered processes.
Only certain data elements for these scenarios are reported in sections 2 through 5 of the
RMP. Section 2 is for toxic substance worst-case scenario data, Section 3 is for toxic alternative
release scenario data, Section 4 is for flammable substance worst-case scenario data, and Section
5 is for flammable alternative release scenario data. (See Table A-l for a list of the data elements
and see the Sample RMP below.)
Table A-l - Data Reported in OCA Sections of an RMP
RMP Sections
Data Elements:
2.1,2.2,3.1,3.2,
4.1, 5.1
Chemical name, percent concentration, and physical state
2.3,3.3,4.2,5.2
Dispersion model used to conduct the analysis (e.g. lookup table, RMP*Comp software)
2.4, 3.4, 4.3, 5.3
Release scenario (e.g., gas leak, liquid spill and vaporization, pipe leak, etc.)
4.5, 5.5
Consequence endpoint assumed (e.g., explosion over pressure, radiant heat level)
(flammable scenarios only; toxic endpoints are mandated by rule)
2.5,2.6,2.7,3.5,
3.6, 3.7, 4.4, 5.4
Quantity released, release rate, and release duration
2.8, 3.8
Wind speed (for worst-case, must be 1.5 meters/sec unless facility has other data)
2.9, 3.9
Atmospheric stability class (for worst-case, must be most stable [F] unless facility has other
data)
2.10,3.10
Topography of area surrounding the process or facility (urban or rural)
2.11,3.11,4.6,
5.6
Distance in miles to either the toxic or flammable endpoint
2.12,3.12,4.7,
5.7
Estimated residential population within the endpoint distance
2.13,3.13,4.8,
5.8
Public receptors (e.g., schools, residences, recreation areas, etc.) within the endpoint
distance
A-l 1
-------
2.14,3.14,4.9,
5.9
Environmental receptors (e.g., national or state parks, etc.) within the endpoint distance
2.15,3.15,4.10,
5.10
Passive mitigation considered (i.e., equipment that functions without human, mechanical,
or energy input that is designed to limit a release)
3.16, 5.11
Active mitigation considered (alternative scenarios only)
...
Graphics file name (optional). Facilities may include a map or other graphic to illustrate a
release scenario
10. What data does not appear in OCA information?
It is important to note what data does not appear in OCA information (or any other
portion of an RMP). OCA information does not contain the following:
• The number of people that would be killed by a worst-case scenario. While the OCA
information contains estimates of affected populations inside worst case and alternative
release scenario circles, these are not estimates of the number of fatalities or injuries that
would occur following such a scenario. As described above, the population reported for a
worst-case scenario includes the total population inside a circle whose radius is equal to
the distance to a particular endpoint. The endpoint is the concentration of toxic substance
in a cloud (for toxic substances) or the radiant heat or overpressure (flammable
substances) beyond which someone could be exposed for a short time and suffer no
serious irreversible injury. However, since toxic gas clouds generally travel in the
direction of the prevailing wind, they form a long, narrow plume, which covers only a
relatively small fraction of the worst-case circle. Therefore, only a small fraction of the
people inside that circle would actually be affected by the cloud. Furthermore, since the
endpoint of the cloud is much lower than the fatal toxic concentration, the number of
fatalities resulting from the release would be smaller yet. For flammable gas scenarios, the
blast effects would likely be felt in all directions from the source, so all people inside the
circle could feel its effects. However, the endpoint for worst-case blast effects is 1 psi
over pressure, which is also far below the level that would cause fatalities. Finally, the
population count assumes all persons remain in place within the circle and are fully
exposed for the time necessary to generate an effect; in an actual emergency, people
shelter-in-place or evacuate and do not receive sufficient exposure to generate any ill
effects.
• How to cause a worst-case scenario. Although the OCA indicates the scenario used to
generate a release (e.g. vessel failure or hose rupture), the OCA contains no information
describing how to make a worst-case or alternative scenario actually happen. OCA
scenarios, and particularly worst-case scenarios, could generally only occur under a
combination of very unusual conditions. Virtually no single event, such as detonation of a
chemical explosive, could initiate a worst-case scenario. Studies of severe chemical plant
A-12
-------
accidents have shown that such accidents have usually resulted from the confluence of
multiple abnormal events or conditions in process or management systems (Loss
Prevention in the Process Industries, Second Edition, F.P. Lees, Butterworth-Heinemann
Publishing, 1996, pp 2/2-3). In Bhopal, four separate safety systems, any one of which
would have prevented the accident, had been disabled prior to the accident, and a fifth
failed to operate properly (Safety in the Chemical Industry, Lessons from Major
Disasters, E.A. Stallworthy and O.P. Kharbanda, G.P. Publishing, Columbia, MD, 1988,
pp 99-100). .
Investigations also show that single explosions rarely initiate worst-case toxic or
flammable gas accidents. When viewed from a common-sense standpoint, the reason
becomes clear. An explosion at a chemical plant often immediately starts a fire. If the
initiating explosion also results in the release of a toxic substance, the ensuing fire will
generally either partially or completely combust the toxin, reducing its toxicity while
simultaneously dispersing it upward and away from surrounding populations. In the case
of flammable materials, an initiating explosion will generally immediately ignite any
flammable material released, causing a large fire, but actually preventing a much more
severe vapor cloud explosion (the worst-case flammable accident). Vapor cloud
explosions require that an extended release of flammable gas occur into a confined area
prior to gas ignition. If the gas is immediately ignited upon release, a fire occurs, but not
an explosion. Numerous experimental programs devoted to the study of vapor cloud
explosions have shown that such explosions are difficult to reproduce, even under
carefully controlled conditions (Guidelines for Evaluating the Characteristics of Vapor
Cloud Explosions, Flash Fires, and BLEVEs, CCPS/AIChE, 1994, pp 70-75). Even
accidental explosions at facilities that store, transport, and manufacture explosives have
usually resulted in little damage (Stallworthy and Kharbanda, pp 67-68).
Finally, even if someone were able to trigger a vessel failure for example, the
worst-case scenario assumes that all of the substance is released and becomes airborne
within 10 minutes under extreme weather conditions. No one has any control over the
weather and there is no guarantee that the vessel could be failed such that all of the
chemical contained therein would be released quickly enough.
• The specific location of toxic or flammable substances. OCA information contains no
information on site layout in general or the specific location of tanks, pipes, or vessels that
contain toxic or flammable materials.
Other significant data which are not found in OCA information include:
The location and design of release mitigation systems;
• Operating procedures for toxic or flammable material processes or release mitigation
systems, or their set points or operating parameters;
• Design or construction information for any process equipment;
• Actual or prevailing meteorological conditions;
A-13
-------
• Site security features or plans; and
• Site staffing plan or operations schedules
11. What OCA information is available to the public right now?
While CSISSFRRA does not give the public a right of access to the EPA OCA
information database prior to the promulgation of regulations, it does provide several mechanisms
by which the public may nevertheless get access to limited amounts of OCA information before
the regulations are issued; these mechanisms are described in detail in Appendix B.
12. What OCA data elements (i.e., OCA data not in RMP or EPA database format) are
available to the public now?
The public has access now to at least some of the OCA data elements in the RMP
executive summary. Facilities are required by the RMP regulations to include at least a brief
description of their OCA in the executive summary of their RMPs; many facilities have included at
least some of the OCA data elements in their summaries and those summaries are already posted
on the Internet along with the rest of their RMPs (minus sections 2-5). Facilities have wide
latitude to decide how much OCA information to provide in their executive summary. Check the
Sample RMP in section 15 below.
13. What OCA "building block" data does the public have access to now?
Other publicly available information can be gathered and analyzed to provide information
similar to some OCA data elements. For example, an important component of the worst-case
release scenario is the quantity of the toxic or flammable substance in the largest vessel. The
chemical name and quantity of that chemical on-site is available in the registration section of the
RMP for each process. However, for a large facility with several listed substances and/or covered
processes, the public won't know which chemical or process was considered for the potential
worst-case or alternative release or how much of the chemical was expected to be released. Also,
the registration information identifies the quantity of the chemical in the process; this may or may
not be the same as the quantity stored in the largest vessel. Storage quantity is also available
through TRI and other databases, but those data sources have the problems described above and
more (e.g., TRI reports the total amount of a chemical on-site, not in a process)
If the quantity of the regulated substance in the registration information, or as provided by
TRI, are used, it is likely that the scenario would over-estimate the consequence distance since the
maximum quantity in a process or on-site is being used in lieu of the maximum amount stored in a
single vessel. Also, the analysis would not be able to determine or account for the existence of
passive mitigation devices.
For worst-case scenario, the public has access to an EPA calculation model which most
facilities have used to perform their analysis; however, without key inputs (such as those
described above), people may get different (and potentially more alarming) results.
A-14
-------
For alternative scenarios, the facility has wide latitude in its choice of inputs, so without
key inputs, a member of the public attempting to reproduce a facility's OCA will likely get
different results.
The potentially affected population is a function of the consequence distance, so whatever
errors are made in determining this distance will also affect the public's calculation of population
affected; census and other data for determining the population of a given area are publicly
available.
The prevention program portion of RMPs lists all the types of mitigation measures used by
a facility in a process, but without OCA-specific data, the public won't know which measures
were involved in calculation of consequence distance or population affected.
Facility siting information can easily be found in telephone directories (e.g. yellow pages).
In particular, sources of the yellow pages on the Internet are also linked to a map that shows the
exact location of facilities. Also, siting and product information may be found in financial reports
that are available on-line from the Securities and Exchange Commission's EDGAR system.
14. What Other Consequence Information is available to the Public?
Other information that may provide tlie public with information about the actual or
potential consequences of accidental chemical releases include:
1. The 5-year accident history section of RMPs.
2. The accident prevention program sections of RMPs (including the hazard assessment).
3. The accidental release information program is an EPA database that provides detailed
information on the consequences and causes of a selected number of accidental releases.
4. The Emergency Release Notification System (ERNS) is another EPA database that
provides emergency notification information on any release of which the National
Response Center has been notified. This information is provided to the extent it is known
at the time of the release.
15. Sample RMP (fictitious)
Below is a sample of a complete RMP including OCA information (fictitious) and how it
might appear as printed from RMP*Info:
Facility Name' General Pulp & Paper
EPA ID. 1000 0010 1922
Section 1. Registration Information
A-15
-------
1.1 Source Identification: Facility ID: 12345
a. Facility Name: General Pulp & Paper
b. Parent Company #1 Name:
c. Parent Company #2 Name:
1.2 EPA Facility Identifier: 1000 0010 1922
1.3 Other EPA Systems Facility ID: ORD004201977
1.4 Dun and Bradstreet Numbers (DUNS):
a. Facility DUNS: 001201977
b. Parent Company #1 DUNS:
c. Parent Company #2 DUNS:
1.5 Facility Location Address:
a. Street 1: 238 Frontage Road
b. Street 2:
c. City: Odenton d. State: MD
f. County: HOWARD
Facility Latitude and Longitude:
g. Lat. (deg min sec):
e. Zip:
3911150
45.187500
11
PG
h. Long, (deg min sec):
h. Long, (decimal degs.):
Interpolation - Map
Plant Entrance (General)
f. State: MD
d. Street 2:
9- Zip:
21873
g. Lat. (decimal degs.):
i. Lat/Long Method:
j. Lat/Long Description:
1.6 Owner or Operator:
a. Name: General Pulp & Paper
b. Phone: (410)777-1234
Mailing address:
c. Street 1: P.O Box 1234
e. City: Odenton
1.7 Name and title of person or position responsible for part 68 (RMP) implementation:
a. Name of person: John Jones
b. Title of person or position: Plant Manager
1.8 Emergency contact:
a. Name: Mary Smith
b. Title: Chemical Engineer
c. Phone: (410)875-2871
d. 24-hour phone: (410)875-4000
e. Ext. or PIN
a. Facility or Parent Company E-Mail Address:
b. Facility Public Contact Phone:
c. Facility or Parent Company WWW Homepage Address:
1.10 LEPC: Howard County LEPC
1.11 Number of full time employees on site: 538
1.12 Covered by:
a. OSHA PSM: Yes
b. EPCRA 302: Yes
c.CAA Title V: Yes Air operating permit ID: 06-2251
1.13 OSHA Star or Merit Ranking: No
1.14 Last Safety Inspection (by an External Agency) Date: 08/21/1998
21873
-0765010.0
-076.8350
A-16
-------
1.15 Last Safety Inspection Performed by an External Agency:
1.16 Will this RMP involve predictive filing?: No
OSHA
Reporting Center and RMP'Maintain Fields
Submission Method: RMP'Submit
Submission Type: F
Receipt Date:
Postmark Date:
Completeness Check Date:
Error Report Date:
De-registration Date:
De-reglstratlon Effective Date:
Anniversary Date:
Section 1.17 Process(es)
06/22/1999
06/18/1999
07/10/1999
a. Process ID: 89876
b. NAICS Code
32211 Pulp Mills
c. Process Chemicals
c.1 Process Chemical (ID I Name)
19507 Chlorine
Program Level
Program Level
a. Process ID: 89877
b. NAICS Code
32211 Pulp Mills
c. Process Chemicals
c.1 Process Chemical (ID I Name)
19508 Chlorine dioxide [Chlorine oxide (CI02)]
Certification Received: Yes
CBI Substantiation Letter: No
CBI Unsanitized Version: No
Electronic Waiver Present: No
Attachments Received: No
Graphic File Received: No
RMP Complete: Yes
CBI Flag: No
Chlorine System
c.2 CAS Nr.
7782-50-5
Chlorine Dioxide
c.2 CAS Nr.
10049-04-4
Section 2. Toxics: Worst Case
Toxics: Worst Case ID: 87654
2.1 a. Chemical Name: Chlorine
b. Percent Weight of Chemical (if In a mixture):
2.2 Physical State:
2.3 Model used:
2.4 Scenario:
2.5 Quantity released:
2.6 Release rate:
2.7 Release duration:
2.8 Wind speed:
2.9 Atmospheric Stability Class:
2.10 Topography: Rural
2.11 Distance to Endpoint:
Both gas and liquid
DEGADIS
Toxic gas release
180,000 lbs
18,000.0 Ibs/mln
10.0 mlns
1.5 m/sec
10 60 mi
2.12 Estimated Residential population within distance to endpoint:
156,567
A-17
-------
2.13 Public receptors within distance to endpoint:
a. Schools: Yes d. Prisons/Correction facilities: Yes
b. Residences: Yes e. Recreation areas: Yes
c. Hospitals: Yes f. Major commercial, office or, industrial areas: Yes
g. Other (Specify):
2.14 Environmental receptors within distance to endpoint:
a. National or state parks, forests, or monuments: Yes
b. Officially designated wildlife sanctuaries, preserves, or refuges: Yes
c. Federal wilderness areas: No
d. Other (Specify):
2.15 Passive mitigation considered:
a. Dikes: No d. Drains: No
b. Enclosures: No e. Sumps: No
c. Berms: No f. Other (Specify):
Toxics: Worst Case ID: 87655
2.1 a. Chemical Name: Chlorine dioxide [Chlorine oxide (CI02)]
b. Percent Weight of Chemical (if in a mixture):
2.2 Physical State: Gas
2.3 Model used: DEGADIS
2.4 Scenario: Toxic gas release
2.5 Quantity released:
19,860
lbs
2.6 Release rate:
1,986.0
Ibs/min
2.7 Release duration:
10.0
mins
2.8 Wind speed:
1.5
ml sec
2.9 Atmospheric Stability Class: F
2.10 Topography: Rural
2.11 Distance to Endpoint: 7.20 mi
2.12 Estimated Residential population within distance to endpoint: 26,240
2.13 Public receptors within distance to endpoint:
a. Schools: Yes d. Prisons/Correction facilities: Yes
b. Residences: Yes e. Recreation areas: Yes
c. Hospitals: Yes f. Major commercial, office or, industrial areas: Yes
g. Other (Specify):
2.14 Environmental receptors within distance to endpoint:
a. National or state parks, forests, or monuments: No
b. Officially designated wildlife sanctuaries, preserves, or refuges: Yes
c. Federal wilderness areas: No
d. Other (Specify):
a. Dikes: No d. Drains: No
b. Enclosures: No e. Sumps: No
c. Berms: No f. Other (Specify):
A-18
-------
Section 3. Toxics: Alternative Release
Toxics: Alternative Release ID: 98765
3.1 a. Chemical Name: Chlorine
b. Percent Weight of Chemical (If in a mixture):
3.2 Physical State: Both gas and liquid
3.3 Model: DEGADIS
3.4 Scenario: Pipe leak
3.5 Quantity released: 310 lbs
3.6 Release rate: 1550 0 Ibs/min
3.7 Release duration: 0.2 mins
3.8 Wind speed: 3.0 m/sec
3.9 Atmospheric Stability Class: D
3.10 Topography: Rural
3.11 Distance to Endpolnt: 1.80 mi
3.12 Estimated Residential population within distance to endpolnt: 3,930
3.13 Public receptors within distance to endpolnt:
a. Schools: Yes d. Prisons/Correction facilities:
b. Residences: Yes e. Recreation areas:
c. Hospitals: No f. Major commercial, office, or industrial areas:
g. Other (Specify):
3.14 Environmental receptors within distance to endpolnt:
a. National or state parks, forests, or monuments: No
b. Officially designated wildlife sanctuaries, preserves, or refuges: No
c. Federal wilderness areas: No
d. Other (Specify):
3.15 Passive mitigation considered:
a. Dikes: No
b. Enclosures: No
c. Berms: No
3.16 Active mitigation considered:
a. Sprinkler systems: No
b. Deluge system: No
c. Water curtain: No
d. Neutralization: No
e. Excess flow valve:
NO
Yes
Yes
d. Drains: No
e. Sumps: No
f. Other (Specify):
f. Flares:
g. Scrubbers:
h. Emergency shutdown systems:
i. Other (Specify):
No
No
Yes
Toxics: Alternative Release ID: 98766
3.1 a. Chemical Name: Chlorine dioxide [Chlorine oxide (CI02)]
b. Percent Weight of Chemical (If In a mixture):
3.2 Physical State:
3.3 Model:
3.4 Scenario:
3.5 Quantity released:
3.6 Release rate:
3.7 Release duration:
Gas
DEGADIS
Pipe leak
1,510 lbs
80.0 Ibs/min
20.0 mins
A-19
-------
3.8 Wind speed: 3.0 m/sec
3.9 Atmospheric Stability Class: D
3.10 Topography: Rural
3.11 Distance to Endpoint: 2 13 mi
3.12 Estimated Residential population within distance to endpoint: 4,660
3.13 Public receptors within distance to endpoint:
a. Schools: Yes d. Prisons/Correction facilities:
b. Residences: Yes e. Recreation areas:
c. Hospitals: Yes f. Major commercial, office, or industrial areas:
g. Other (Specify):
3.14 Environmental receptors within distance to endpoint:
a. National or state parks, forests, or monuments: No
b. Officially designated wildlife sanctuaries, preserves, or refuges: No
c. Federal wilderness areas: No
d. Other (Specify):
3.15 Passive mitigation considered:
a. Dikes: No
b. Enclosures: No
c. Berms: No
3.16 Active mitigation considered:
a. Sprinkler systems: No
b. Deluge system: No
c. Water curtain: No
d. Neutralization: No
e. Excess flow valve: No
No
Yes
Yes
d. Drains: Yes
e. Sumps: Yes
f. Other (Specify):
f. Flares:
g. Scrubbers:
h. Emergency shutdown systems:
I. Other (Specify):
No
No
No
Section 4. Flammables: Worst Case — No Data To Report
Section 5. Flammables: Alternative Release — No Data To Report
Section 6. Accident History
Accident History ID: 3567
6.1 Date of accident: 12/03/1996 6.2 Time accident began(HHMM):
6.3 NAICS Code of process involved: 32211
6.4 Release duration: 000 Hours (HHH) 20 Minutes (MM)
6.5 Chemical(s):
a. Chemical Name CAS Number
Chlorine 7782-50-5
1330
b. Quantity c. %
Released (lbs) Weight
6.6 Release event:
a. Gas release:
b. Liquid spill/evaporation:
c. Fire:
6.7 Release source:
Yes a. Storage vessel:
No b. Piping:
No c. Process vessel:
No e. Valve: No
No f. Pump: No
No g. Joint: No
A-20
-------
d. Explosion:
No
d. Transfer hose: Yes h. Other (Specify):
6.8 Weather conditions at time of event (if known):
a. Wind speed: Units: meters/second
b. Temperature: Degrees Fahrenheit
c. Atmospheric Stability Class:
d. Precipitation present: No
e. Unknown weather conditions: Yes
6.9 On-site Impacts:
Employees or contractors: Public responders:
a. Deaths
b. Injuries
c. Property damage ($):
6.10 Known Off-site impacts:
a. Deaths:
b. Hospitalization:
c. Other medical treatments:
2,500
Direction:
0 d. Evacuated:
0 e. Sheltered-in-place:
0 f. Property Damage ($):
Public:
0
g. Environmental damage:
1. Fish or Animal Kills: No
2. Tree, lawn, shrub, or crop damage: No
3. Water contamination:
No
4. Soil contamination:
No
5. Other (specify):
Initiating event: a
Equipment Failure
Contributing factors:
a. Equipment failure:
Yes
g. Maintenance activity/inactivity:
No
b. Human error:
Yes
h. Process design failure:
No
c. Improper procedures:
No
i. Unsuitable equipment:
No
d. Overpressurizatlon:
No
j. Unusual weather condition:
No
e. Upset condition:
No
k. Management error:
No
f. By-pass condition:
No
I. Other (Specify):
6.13 Offsite responders notified: Notified and Responded
6.14 Changes introduced as a result of the accident:
a. Improved or upgraded equipment:
Yes
g. Revised emergency response plan:
No
b. Revised maintenance:
No
h. Changed process:
No
c. Revised training:
No
I. Reduced inventory:
No
d. Revised operating procedures:
Yes
j. None:
No
e. New process controls:
No
k. Other(Speclfy):
f. New mitigation systems:
No
A-21
-------
Section 7. Prevention Program 3
Process ID: 89876 Chlorine System
Prevention Program ID: 8049
Prevention Program Description:
Chlorine Process
7.1 NAICS Code 32211
7.2 Chemicals
Chemical Name
Chlorine
7.3 Date on which the safety information was last reviewed or revised:
7.4 Process Hazard Analysis (PHA):
a. Date of last PHA or PHA update: 03/05/1999
b. The technique used:
06/11/1998
Failure Mode and Effects Analysis:
Fault Tree Analysis:
HAZOP:
No
No
Yes
What If: No
Checklist: No
What If/Checklist: No
Other (Specify):
c. Expected or actual date of completion of all changes from last PHA or PHA update:
d. Major hazards identified:
06/11/1999.
Toxic release:
Yes
Contamination:
Yes
Fire:
Yes
Equipment failure:
Yes
Explosion:
Yes
Loss of cooling, heating, electricity, instrument air:
Yes
Runaway reaction:
No
Earthquake:
Yes
Polymerization:
No
Floods (flood plain):
Yes
Overpressurization:
Yes
Tornado:
No
Corrosion:
Yes
Hurricanes:
No
Overfilling:
Yes
Other (Specify):
e. Process controls in use:
Vents: Yes
Relief valves: Yes
Check valves: Yes
Scrubbers: Yes
Flares: No
Manual shutoffs: Yes
Automatic shutoffs: Yes
Interlocks: Yes
Alarms and procedures: No
Keyed bypass: No
Other (Specify):
f. Mitigation systems in use:
Sprinkler system: No
Dikes: No
Firewalls: No
Blast walls: No
Emergency air supply: No
Emergency power: Yes
Backup pump: No
Grounding equipment: No
Inhibitor addition: No
Rupture disks: Yes
Excess flow device: Yes
Quench system: No
Purge system: Yes
None: No
Water curtain:
Enclosure:
Neutralization:
None:
No
Yes
Yes
No
A-22
-------
Deluge system: No Other (Specify):
g. Monitoring/detection systems in use:
Process area detectors: Yes None: No
Perimeter monitors: No Other (Specify): Video Surveillance
h. Changes since last PHA or PHA update:
Reduction in chemical inventory:
No
Installation of perimeter monitoring systems:
No
Increase In chemical inventory:
No
Installation of mitigation systems:
No
Change process parameters:
No
None recommended:
Yes
Installation of process controls:
No
None:
No
Installation of process detection systems:
No
Other (Specify):
7.5 Date of most recent review or revision of operating procedures: 05/01/1999
7.6 Training:
a. The date of the most recent review or revision of training programs: 05/01/1999
b. The type of training provided:
Classroom: Yes On the Job: Yes Other (Specify):
c. The type of competency testing used:
Written test: Yes Observation: Yes
Oral test: No Demonstration: Yes
Other (Specify):
7.7 Maintenance:
a. The date of the most recent review or revision of maintenance procedures
b. The date of the most recent equipment inspection or test:
c. Equipment most recently inspected or tested : piping
05/01/1999
06/11/1999
7.8 Management of change:
a. The date of the most recent change that triggered management of
change procedures:
b. The date of the most recent review or revision of management of
change procedures:
7.9 The date of the most recent pre-startup review:
7.10 Compliance audits:
a. The date of the most recent compliance audit:
b. Expected date of completion of all changes resulting from the compliance audit:
7.11 Incident investigation:
a. The date of the most recent incident Investigation (if any):
b. Expected or actual date of completion of all changes resulting from the investigation:
7.12 The date of the most recent review or revision of employee participation plans:
7.13 The date of the most recent review or revision of hot work permit procedures:
7.14 The date of the most recent review or revision of contractor safety procedures:
7.15 The date of the most recent evaluation of contractor safety performance:
05/24/1999
3/18/1999
08/22/1998
06/11/1999
06/18/1999
12/04/1996
03/15/1997
03/11/1999
04/13/1999
04/18/1999
04/04/1999
A-23
-------
Process ID: 89877 Chlorine Dioxide
Prevention Program ID: 8050
Prevention Program Description:
Chlorine Dioxide System
7.1 NAICS Code 32211
7.2 Chemicals Chemical Name
Chlorine dioxide [Chlorine oxide (CI02)]
7.3 Date on which the safety information was last reviewed or revised:
7.4 Process Hazard Analysis (PHA):
a. Date of last PHA or PHA update: 02/16/1999
b. The technique used:
06/30/1997
Failure Mode and Effects Analysis:
Fault Tree Analysis:
HAZOP:
No
No
Yes
What If: No
Checklist: No
What if/Checklist: No
Other (Specify):
c. Expected or actual date of completion of all changes from last PHA or PHA update:
d. Major hazards identified:
05/18/1999
Toxic release:
Yes
Contamination:
Yes
Fire:
Yes
Equipment failure:
Yes
Explosion:
Yes
Loss of cooling, heating, electricity, instrument air:
Yes
Runaway reaction:
No
Earthquake:
Yes
Polymerization:
No
Floods (flood plain):
Yes
Overpressurization:
Yes
Tornado:
No
Corrosion:
Yes
Hurricanes:
No
Overfilling:
Yes
Other (Specify):
rocess controls in use:
Vents:
Yes
Emergency air supply:
No
Relief valves:
Yes
Emergency power:
Yes
Check valves:
Yes
Backup pump:
No
Scrubbers:
Yes
Grounding equipment:
No
Flares:
No
Inhibitor addition:
No
Manual shutoffs:
Yes
Rupture disks:
Yes
Automatic shutoffs:
Yes
Excess flow device:
Yes
Interlocks:
Yes
Quench system:
No
Alarms and procedures:
Yes
Purge system:
Yes
Other (Specify):
f. Mitigation systems in use:
Sprinkler system: No
Dikes: No
Fire walls: No
Blast walls: No
Deluge system: No
Water curtain:
Enclosure:
Neutralization:
None:
Other (Specify):
No
Yes
Yes
No
g. Monitoring/detection systems in use:
A-24
-------
Process area detectors: Yes None: No
Perimeter monitors: No Other (Specify):
h. Changes since last PHA or PHA update:
Reduction in chemical inventory:
No
Installation of perimeter monitoring systems:
No
Increase in chemical inventory:
No
Installation of mitigation systems:
No
Change process parameters:
No
None recommended:
Yes
Installation of process controls:
No
None:
No
Installation of process detection systems:
No
Other (Specify):
Keyed bypass: No
None:
No
7.5 Date of most recent review or revision of operating procedures: 04/05/1999
7.6 Training:
a. The date of the most recent review or revision of training programs: 04/20/1999
b. The type of training provided:
Classroom: Yes On the job: Yes Other (Specify):
c. The type of competency testing used:
Written test: Yes Observation: Yes
Oral test: No Demonstration: Yes
Other (Specify):
7.7 Maintenance:
a. The date of the most recent review or revision of maintenance procedures:
b. The date of the most recent equipment inspection or test:
c. Equipment most recently inspected or tested : Chlorine Dioxide Production System
7.8 Management of change:
a. The date of the most recent change that triggered management of
change procedures:
b. The date of the most recent review or revision of management of
change procedures:
7.9 The date of the most recent pre-startup review:
7.10 Compliance audits:
a. The date of the most recent compliance audit:
b. Expected date of completion of all changes resulting from the compliance audit:
7.11 Incident investigation:
a. The date of the most recent incident Investigation (if any):
b. Expected or actual date of completion of all changes resulting from the investigation:
7.12 The date of the most recent review or revision of employee participation plans:
7.13 The date of the most recent review or revision of hot work permit procedures:
7.14 The date of the most recent review or revision of contractor safety procedures:
7.15 The date of the most recent evaluation of contractor safety performance:
Section 8. Prevention Program 2 — No Data To Report
04/14/1999
05/18/1999
02/23/1999
03/18/1999
10/10/1997
06/11/1999
06/18/1999
03/11/1999
04/13/1999
04/18/1999
04/04/1999
A-25
-------
Section 9. Emergency Response
9.1 Written Emergency Response (ER) Plan:
a. Is facility included in written community emergency response plan? Yes
b. Does facility have its own written emergency response plan? Yes
9.2 Does facility's ER plan include specific actions to be taken in
response to accidental releases of regulated substance(s)? Yes
9.3 Does facility's ER plan include procedures for Informing the
public and local agencies responding to accidental releases? Yes
9.4 Does facility's ER plan include information on emergency heath
care? Yes
9.5 Date of most recent review or update of facility's ER plan: 11/30/1998
9.6 Date of most recent ER training for facility's employees: 02/06/1999
9.7 Local agency with which facility's ER plan or response activities are coordinated:
a. Name of agency: Howard County Fire Department
b. Telephone number: (410)-321-7654
9.8 Subject to:
a. OSHA Regulations at 29 CFR 1910.38: Yes
b. OSHA Regulations at 29 CFR 1910.120: Yes
c. Clean Water Act Regulations at 40 CFR 112: Yes
d. RCRA Regulations at 40 CFR 264, 265, and 279.52: Yes
e. OPA-90 Regulations at 40 CFR 112, 33 CFR 154, 49 CFR 194, or 30 CFR 254: Yes
f. State EPCRA Rules/Law: No
Executive Summary
This Risk Management Plan (RMP) is submitted to the U.S. Environmental Protection Agency (U S. EPA) for General Pulp & Paper in
accordance with the requirements of Section 112(r) of the Clean Air Act Amendments of 1990 as codifed in Title 40 of the Code of
Federal Regulations (CFR) Part 68 General Pulp & Paper handles two regulated substances listed in Appendix A of Part 68.
1.1 STATIONARY SOURCE & REGULATED SUBSTANCES HANDLED
General Pulp & Paper owns and operates a pulp and paper mill located in Odenton, Md. The regulated substances handled by this
facility are chlorine and chlorine dioxide, both of which are on the U S EPA's list of regulated toxic substances for CAA section 112(r)
The Odenton plant produces pulp and paper from wood chips and sawdust using the Kraft process Chlorine and chlorine dioxide are
used in a bleaching process to remove lignin from the fibers and to whiten pulp, chlorine is also used to treat process water
Liquid chlorine is stored in rail cars and storage tanks prior to use and fed to a vaporizer Gaseous chlorine is then fed to the process
The maximum quantity of chlorine that stored at this facility is 600,000 pounds.
Chlorine dioxide is generated on site by a process which uses sodium chlorate, methanol, and sulfuric acid as a raw materials. These
raw materials are not regulated under section 112(r) The chlorine dioxide produced in the process is absorbed into water and then
stored as a dilute aqueous solution (10 g/l) The maximum quantity of chlorine dioxide stored at this facility is 35,000 pounds
1 2 ACCIDENTAL RELEASE PREVENTION & EMERGENCY RESPONSE PROGRAMS
General Pulp & Paper prevents chemical accidents using an integrated process safety management system. The plant uses several
A-26
-------
management systems and follows applicable industry and national standards to meet this goal
General Pulp & Paper's chlorine and chlorine dioxide processes are covered by the OSHA Process Safety Management (PSM)
standard (29 CFR 1910 119). General Pulp & Paper adheres strictly to the PSM standard and focuses many of its safety efforts
around PSM. The PSM program requires General Pulp & Paper to take specific efforts to identify and mitigate process hazards and
prevent accidents. The elements of the PSM program are very similar to the accident prevention elements in the EPA risk
management program, which General Pulp and Paper also fully implements.
Although an accidental chemical release is unlikely, General Pulp & Paper prepares for releases and other emergencies. The plant
has developed and implemented a written response plan which is discussed is detail in Section 1.5. General Pulp & Paper employees
routinely practice responding to simulated releases and emergencies, and coordinate with community responders such as the
Odenton Fire Department.
1.3 WORST-CASE & ALTERNATIVE RELEASE SCENARIOS
General Pulp & Paper has constructed a worse-case release scenario and alternate (i.e. more credible) release scenario for each
regulated chemical.
CHLORINE: WORST-CASE SCENARIO
The failure of the largest storage tank (i.e railcar) when filled to the greatest amount allowed would release 180,000 pounds of
chlorine Since the contents of the railcar are under pressure, the release is assumed to be a liquid jet that volatilizes to gas upon
release from the tank. The entire contents of the railcar are assumed to release at a constant rate over a ten minute period.
CHLORINE DIOXIDE WORST-CASE SCENARIO
The failure of our largest chlorine dioxide solution storage tanks would release 238,000 gallons of chlonne dioxide solution, or 19,856
pounds of chlorine dioxide Company policy limits the maximum filling capacity of the large chlorine dioxide storage tanks to 90%; the
238,000 gallon figure is 90% of the physical capacity of the tank. It is assumed that the entire contents of the tank are released and
instantaneously form a pool 1 cm deep The chlorine dioxide volatilization rate from the pool is calculated according to a model based
on an evaporative pool model.
CHLORINE ALTERNATE SCENARIO
A 1" pipe conveys liquid chlorine to the water treatment plant from the chlorine expansion tank This pipe could be ruptured by a
vehicle (e g. forklift) striking the pipe bridge which contains the chlorine pipe. This would release 310 pounds of liquid chlorine that is
assumed to vaporize instantly. The release is estimated to take twelve seconds
CHLORINE DIOXIDE. ALTERNATE SCENARIO
A fiberglass pipe which conveys chlorine dioxide from the large storage tanks to the bleach plant is assumed to be damaged by
mechanical impact during a pump replacement or other maintenance work. A 3" diameter hole is made in the pipe and chlorine dioxide
solution is released The motive force is the gravity head of the tank; it is assumed that the pump is shut off immediately during the
evacuation of the area. The release continues for twenty minutes until a response crew can enter the required protective equipment
and shut off the release. A drain in the vicinity of the pipe is assumed to be able to capture 2 gallons per second of the spill, this is
directed to a gas-tight sump where the spilled material can be collected and treated.
1 4 FIVE YEAR ACCIDENT HISTORY
General Pulp & Paper has had one release of a regulated material that resulted in an injury in the last five years. On December 3,
1996, an employee was injured when exposed to chlorine leaking from a hose. There have been no releases of regulated materials
which have resulted in deaths, significant property damage, or any known offsite deaths, injuries, evacuations, sheltering in place,
property damage, or environmental damage in the last five years.
1.5 EMERGENCY RESPONSE PROGRAM
In addition to the prevention program, General Pulp & Paper has developed and implemented a written emergency response plan to
effectively respond to accidental chemical releases This plan identifies roles for plant personnel in the event of a number of different
scenarios The plan includes specific tasks for key personnel dunng responses, emergency plant shutdown procedures, steps to
contain and handle releases of specific materials, specific information on how to contact community response agencies and the public,
and information on training employees and community responders in safe response techniques. General Pulp & Paper trains regularly
on its emergency plan This training includes plant employees, members of General Pulp & Paper's response team, and community
responders. Training exercises are evaluated, and the plan is updated when deficiencies are identified.
General Pulp & Paper maintains an emergency response team that is trained to respond to many different types of emergencies. The
team is made up of workers from different shifts and is always ready to respond The team regularly conducts response drills, often
A-27
-------
including community responders
1.6 PLANNED CHANGES TO IMPROVE SAFETY
General Pulp & Paper has identified no major unresolved process hazards in the chlonne or chlorine dioxide systems No major
revisions to those processes are currently planned However, General Pulp & Paper follows a policy of continuous process safety
improvement
A-28
-------
APPENDIX B
DISCUSSION OF THE APPLICABLE PROVISIONS OF THE
CHEMICAL SAFETY INFORMATION, SITE SECURITY, AND FUELS
REGULATORY RELIEF ACT (CSISSFRRA)
B-l
-------
APPENDIX B
DISCUSSION OF THE APPLICABLE PROVISIONS OF THE
CHEMICAL SAFETY INFORMATION, SITE SECURITY, AND FUELS
REGULATORY RELIEF ACT (CSISSFRRA)
1. What is CSISSFRRA, and how does it relate to this assessment?
In August of 1999, Congress passed the Chemical Safety Information, Site Security and
Fuels Regulatory Relief Act (CSISSFRRA) to address concerns about potential Internet posting
of a database containing information on the consequences of hypothetical chemical accidents.1
Under a regulatory program required by the Clean Air Act (CAA), facilities handling certain very
hazardous substances must conduct analyses of the off-site consequences of such hypothetical
accidents and report the results (off-site consequence analysis information or OCA information -
see Appendix A) in a plan submitted to the U.S. Environmental Protection Agency (EPA).
CSISSFRRA temporarily exempts OCA information from public disclosure under the CAA and
the Freedom of Information Act (FOIA).
CSISSFRRA also requires the President to assess the increased risk of terrorist and other
criminal activity associated with the posting of OCA information on the Internet; and the
incentives created by public disclosure of OCA information to reduce the risk of accidental
chemical releases. Based on the assessments, the President is to issue regulations governing the
distribution of OCA information in a manner that, in the opinion of the President, minimizes the
likelihood of accidental releases and any increased risk of terrorist activity associated with Internet
posting of OCA information and the likelihood of harm to public health and welfare. The
President has delegated to the Department of Justice (DOJ) the responsibility of assessing the
increased risk of terrorist and criminal activity and to EPA the responsibility of assessing the
incentives for reduction in chemical accidents created by public disclosure of OCA information.
On January 27, 2000, the President provided joint delegation to DOJ and EPA to promulgate the
regulations, after review and approval by the Office of Management and Budget.
As noted above, CSISSFRRA exempts "off-site consequence analysis information" from
FOIA for at least one year while the President assesses the criminal risks of posting the
information on the Internet and the chemical safety benefits of providing public access to the
information and then issues regulations governing distribution of the information based on the
assessments (section 112(r)(7)(H)(ii)). CSISSFRRA defines "off-site consequence analysis
information" (OCA information) as the OCA sections of any RMP submitted to EPA and any
electronic database EPA creates from those sections (section 112(r)(7)(H)(l)(III)). It expressly
'CSISSFRRA also contains provisions which prohibit EPA from regulating flammable substances
under the Risk Management Program when those substances are used as fuel or held for sale as fuel at a
retail facility. The fuel provisions of CSISSFRRA are not discussed in this Appendix.
B-2
-------
excludes an RMP's Executive Summary, which is required to include at least a brief description of
the submitting facility's OCA.
2. How can the public gain access to OCA information under CSISSFRRA?
CSISSFRRA provides the public with other means of access to the data reported in the
OCA sections of RMPs, and even to the OCA sections themselves, before and/or after the federal
government conducts its assessments and rulemaking. As noted above, RMP Executive
Summaries are not covered by its restrictions, and facilities are required to provide at least a brief
description of their OCAs in their Executive Summaries. The summaries are already available on
the Internet through several web sites. A random sampling of the summaries indicates that the
amount of OCA information reported varies from facility to facility; some facilities provided
nearly complete information while others provided little. (EPA's rule does not define a "brief
description," leaving facilities to make reasonable decisions as to what information to include.) In
addition, CSISSFRRA requires virtually all covered facilities to conduct a public meeting or post
a public notice by February 5, 2000, that summarizes their OCA information (CSISSFRRA
section 4). To date, the Federal Bureau of Investigation has received notification from about
5,000 facilities that they have complied with this requirement.
CSISSFRRA also does not prevent facilities from releasing their OCA information to the
public without restriction, and once a facility has so released its OCA information, covered
persons may do so as well (section 112(r)(7)(H)(v)(III)). To date, EPA has received notification
from over 900 facilities that they have released their OCA information without restriction.
CSISSFRRA further provides that states which collect data on off-site consequences, even data
identical in content and format to OCA information, are not precluded from releasing it to the
public (section 112(r)(7)(H)(x)(II)). Several states have in place state laws requiring the
collection of similar or identical data.
CSISSFRRA guarantees public access to OCA information itself (i.e., the OCA sections
of RMPs and EPA's database created from those sections) in several ways. First, it requires EPA
to provide the public with OCA information without information concerning the identity and
location of the facilities reporting the information (section 112(r)(7)(H)(iv)). EPA is consulting
with other federal agencies and stakeholders to implement this provision. Second, CSISSFRRA
requires EPA, in consultation with DOJ and other agencies, to establish a "read-only information
technology system" that "provides for the availability to the public of [OCA information] by
means of a central data base under the control of the Federal Government that contains
information that users may read, but that provides no means by which an electronic or mechanical
copy of the information may be made" (section 112(r)(7)(H)(viii)). EPA is working with other
federal agencies to identify the best methods for development of this read-only system. Third,
CSISSFRRA requires EPA, in consultation with DOJ, to make OCA information available to
"qualified researchers" by means of a system that does not allow researchers who receive the
information to disseminate it (section 112(r)(7)(H)(vii)). Finally, CSISSFRRA provides that, at a
minimum, the regulations based on the assessment must "allow access by any member of the
public to paper copies of [OCA] information for a limited number of stationary sources located
B-3
-------
anywhere in the United States, without any geographical restriction." In short, any member of the
public will be able to have access to paper copies of OCA information for at least some number of
facilities.
3. Who are "covered persons" and how does CSISSFRRA affect them?
Both before and after the regulations are issued, CSISSFRRA guarantees "covered
persons" access to OCA information for their "official use" (see section 112(r)(7)(H)(iv) and
(ii)(cc)-(ee)).
The SERC/LEPC category includes
members of 50 State Emergency Response
Commissions and about 3,400 Local
Emergency Planning Committees created
under EPCRA. Members of these
commissions and committees can include
members of public, the media, and industry, as
well as representatives of emergency
responders such as fire and police departments
(EPCRA section 301(c)). Considering that
covered persons include all of the entities
above, there are potentially well over 1 million
covered persons.
Covered persons are guaranteed
access to OCA information (in either RMP or database format) for any or all covered facilities for
"official use" both now and in the future (i.e., under the regulations); members of the public have
no right of access prior to the regulations, but are guaranteed access to at least paper copies of
the OCA sections of RMPs for a limited number of facilities under the regulations and to a read-
only database of OCA information that EPA is to establish in consultation with other federal
agencies.
While CSISSFRRA guarantees covered persons access to OCA information, it prohibits
them from disclosing the information to the public except as authorized by the statute or the
regulations issued under it (section 112(r)(7)(H)(v)). Any covered person who violates the
prohibition is subject to criminal penalties of up to SI,000,000 per year. At the same time,
CSISSFRRA states that it "does not restrict the dissemination of off-site consequence analysis
information by any covered person in any manner or form except in the form of a [RMP] or an
electronic data base created by [EPA] from off-site consequence analysis information" (section
112(r)(7)(H)(xii)(Il)). CSISSFRRA's prohibition on public disclosure is thus narrow. It applies
to the OCA sections of the RMP (sections 2-5 of the RMP form) and any database created by
EPA from those sections, but it does not apply to the information reported in those sections when
provided in a different format, or to the information provided in Executive Summaries. Covered
persons are consequently allowed to communicate the information in the OCA sections of RMPs
"Covered persons" include:
• Federal government officers and
employees and their contractors
• State government officers and
employees and their contractors
Local government officers and
employees and their contractors
• SERC and LEPC members and their
contractors
• State and local police
• Paid and volunteer firefighters
• Other emergency responders
B-4
-------
to the public so long as they do so in a way that does not replicate those sections of the RMP or
EPA's database.
B-5
-------
This page intentionally left blank.
B-6
-------
APPENDIX C
USES OF RIGHT-TO-KNOW INFORMATION
C-l
-------
APPENDIX C
USES OF RIGHT-TO-KNOW INFORMATION
This appendix summarizes over 40 documented cases in which right-to-know information
was used to improve conditions for communities. The first section highlights uses of EPCRA
(Emergency Planning and Community Right-to-Know Act) data. Most of these uses rely on
Toxics Release Inventory (TRI) data, which are available electronically from a central source, but
there are some uses of other EPCRA data as well. The second section describes uses of
information from environmental and other programs outside of EPCRA to reduce risk in some
form.
These uses of right-to-know information shed light on OCA information - on how and
how much it would likely be used under various disclosure schemes. For example, usage of easily
available TRI data seems to be much higher than usage of other EPCRA data, which are more
difficult to obtain. In addition, there is a wide variety of users and of uses of right-to-know
information. Most cases involve multiple segments of the public, with the primary actors ranging
from community, public interest, and environmental organizations, to news media, government,
industry, unions, and research organizations. In the many different uses cited, the most common
outcomes are release/risk reduction, chemical substitution, increased communication, "good
neighbor" agreements, laws/regulations, and improved emergency planning.
A. USES OF EPCRA INFORMATION
IBM Plant Agrees to Eliminate Use of CFCs - San Jose, CA - 1989: An analysis of 1987
Toxic Release Inventory (TRI) data by the Citizens For a Better Environment showed that IBM's
Silicon Valley plant was the largest emitter of chlorofluorocarbons (CFCs) in the state of
California. The Silicon Valley Toxics Coalition organized local labor and environmental groups
to pressure IBM for changes. In July 1989, a front page story in USA Today named IBM's
Silicon Valley plant as the third largest (by volume) emitter of CFCs in the nation. Continuing to
build support with newsletters, meetings, and contact with local enforcement agencies, the
coalition proposed a "good neighbor" agreement that called for IBM to phase-out all use of
CFCs. In September 1989, under the weight of increasing public pressure and negative publicity,
IBM senior management announced a proposal to eliminate all use of CFCs in their products and
processes by 1993. Not only did IBM switch to a safe substitute, but it asked their suppliers to
do the same.1
TRI Data Used to Compile a "Green Index" of Biggest Manufacturers - US - 1993: Fortune
Magazine compiled a "green index" of America's biggest manufacturers using TRI data as a
central element. Fortune examined the environmental records of a number of companies,
developing a relative ranking system that scored the companies from zero to 10 in 20 different
performance categories, such as the amount of toxic emissions per dollar value of sales, and their
percent reduction in toxic emissions.2
C-2
-------
Companies Switch to Less Dangerous Chemicals - Cuyahoga County, OH - 1990: Using
data made available through EPCRA, the Cuyahoga County Local Emergency Planning
Committee (LEPC) conducted a hazard analysis of nearly 300 facilities that handle hazardous
materials. A vulnerability zone was mapped for each facility, marking the area surrounding a
facility that would be effected by a toxic chemical release. Each map was made available in the
local public libraries. As a result of this heightened interest in the safety of surrounding
neighborhoods, Cleveland's largest sewage treatment plant decided to eliminate a 55-ton railroad
tank car of chlorine from its operations. The 1990 Annual Report of the Northeast Ohio Regional
Sewer District announced the change from chlorine gas to liquid sodium hypochlorite, a safer
disinfectant. One plant manager credited right-to-know with increasing awareness and a re-
examination of chemical hazards that had been accepted as routine for years.3
Good-Neighbor Agreement Successful - Berlin, NJ - Using right-to-know data, the New Jersey
Coalition Against Toxics asked five local facilities for the opportunity to inspect the plants for
toxic hazards to workers and the community. Dynasil Corporation of America was the first to
respond. An inspection team, made up of the local fire chief, members of the local emergency
planning committee (LEPC), several neighbors, and two technical consultants toured the facility
and made recommendations as to how Dynasil could improve worker safety and prevent a toxic
disaster. The President of Dynasil made a company commitment to implement the LEPC
suggestions and did so within a month. This cooperation represented one of the first good-
neighbor agreements in the nation.4
Pollution Prevention Through Worker-Management Agreements - New York - The Citizens
Environmental Coalition (CEC), a statewide citizens advocacy organization, uses TRI data for a
number of citizen guides, fact sheets, and information packets. One of the most successful
applications of TRI data by CEC involved a series of workshops that use these data to familiarize
employees with hazards in the workplace. Many workers who attended the workshops are
unaware of much of the TRI data and surprised at the emissions reported by the plants in which
they work. However, workers and management have been able to open dialogues, even leading
to emissions reductions. For example, Harrison Radiator in Lockport stopped using a number of
hazardous chemical solvents because of pressure from workers. In addition, Kodak reduced
emissions from 24 to 14 million pounds, in part as a result of the CEC workshops.5
True Flexibility of TRI Data Shown Through Novel Uses - 1997: An ever-increasing diversity
of uses are being found for TRI data. Insurance companies, stock analysts, house hunters,
epidemiologists, journalists, and all those rating America's best cities are finding the TRI data
valuable. One can now draw correlations, for example, between cancer rates and the amount of
carcinogen releases in a state or local community. In addition, the Detroit News was one of the
first organizations to use the TRI data to examine the "environmental justice" debate, suggesting
that big polluters tend to be located in low-income communities.6
Right-to-Know Empowers Citizens - Contra Costa County, CA - 1989: In February 1989,
Communities for a Better Environment (CBE), a California non-profit, used right-to-know data to
develop a report highlighting the threat of a toxic chemical incident in Contra Costa County. The
C-3
-------
report identified sixty-five companies in the area that collectively stored 140 million pounds of
highly hazardous chemicals. CBE created and distributed a leaflet that summarized TRJ data and
described the location of the storage areas and amount of each chemical stored throughout the
county. CBE also criticized the County Health Department for failing to ask local facilities for
Risk Management Prevention Plans (RMPP), as required by state law.
Following an explosion and fire at a Chevron facility in April 1989, the CBE and the West County
Toxics Coalition wrote to the County Health Department requesting that RMPPs be required
from chemical companies. Armed with data from their right-to-know analysis, the Coalition and
CBE appeared before the County Board of Supervisors and demanded the RMPPs be required for
public safety. The Board agreed, and the first RMPP requests went out on December 1, 1989.7
EPCRA Leads to Decreasing Chemical Hazards in Florida Communities - Florida - 1999:
EPCRA has led to a number of changes in Florida communities that have lowered the chemical-
related risks to local citizens. The requirement to perform a hazard analysis under EPCRA
section 302 led several facilities to work more closely with the local fire department to minimize
or eliminate the risk of a spill. A number of water treatment facilities in the Mel rose/Keystone
Heights area switched from chlorine to a hypochlorite solution, especially important in smaller
communities with limited hazmat response capabilities. Many swimming pools in cities such as
Gainesville now limit the total chlorine that can be stored on-site at any given time. Facility
reviews by the local emergency planning committees (LEPCs) have also spurred the elimination of
pressurized regulators on chlorine systems and a change towards safer vacuum systems, which are
now standard. LEPCs, such as the North Central Florida LEPC, are now reviewing emergency
response information provided in facility risk management plans, which will supplement site-
specific information already included in the LEPC emergency plan.8
TRI Catalyst for Emission Reductions - North Carolina - 1990: This article claims that the
release of TRJ data to the public was the catalyst that led to passage of the first air toxics
regulations in North Carolina. After the NC Environmental Defense Fund announced that
companies had legally released over 100 million pounds of toxic substances in 1987, Governor
James G. Martin backed the regulations, which took effect in May 1990. These control-oriented
standards, set by the Environmental Management Commission, required hundreds of industries to
reduce emissions of 105 toxic air pollutants.9
Business Community Begins to See Economic Benefits Related to TRI - 1996: Since 1988,
national environmental groups have been using TRI data to identify the top polluters in the United
States. Economic incentives pushed major corporations to quickly reduce emissions to avoid
developing a reputation as a major polluter. Armed with TIU data, local citizen groups are now
able to document their concerns and force companies to address the risk of chemical spills.
Despite initial reservations, many business sectors now view the disclosure of TRI and EPCRA
information as important for public outreach and monitoring of performance. Companies that
previously may have been unaware of the extent of their releases are motivated to reduce
emissions that can now be accurately tracked by citizens groups.
C-4
-------
Since TRI's beginnings, an increasing number of businesses are recognizing the positive link
between public accountability and business performance. In recent years, more than 100
companies have begun issuing annual environmental reports to investors, communities,
environmental groups, and government. The reports, based primarily on TRI data, describe
environmental goals, achievements, and setbacks.10
NICs Scorecards™ Help Explain TRI Data - Charleston, WV - 1994: The National Institute
for Chemical Studies (NICs) has summarized TRI data in a more accessible format to help
improve the emergency preparedness efforts of state emergency response commissions (SERCs)
and local emergency planning committees (LEPCs). Because TRI data can help communities
better understand and manage environmental risks, NICs developed Scorecards™ , customizable
for each locale, that offer the SERCs and LEPCs an excellent avenue for the public dissemination
of toxic release information. The scorecards compress and interpret TRI data and allow
companies to explain the details behind those numbers, facilitating better communication and
understanding."
Labor Union and Community Groups Unite, Company to Reduce Emissions and Use -
Northfield, MN - 1990: The naming of Sheldahl Inc. as one of the nations leading emitters of
airborne carcinogens coincided with contract negotiations between Sheldahl and the
Amalgamated Clothing and Textile Workers Union (ACTWU). The announcement also led to the
formation of two new citizens groups advocating pollution reduction and public health. The
union had long been trying to reduce worker exposure to methylene chloride, and the increased
concern for public safety added new weight to the negotiations. The union included
environmental issues in their new contract negotiations. In an effort to ensure that public
concerns did not shut the plant down, the union insisted that local citizens groups be present for
the pollution negotiations. The resulting agreement phased out the use of methylene chloride in
production by switching to a non-toxic substitute and called for a 90 percent emissions reduction
by 1993.12
TRI Triggers Emissions Monitoring Network and Reductions - Rochester, NY - 1992:
Spurred by 1989 TRI data indicating that its Rochester, NY facility ranked second in the United
States for emissions of dichloromethane (DCM), Eastman Kodak pledged to cut DCM emissions
by 70 percent by 1995. To track its progress in meeting these goals, Kodak stepped up emissions
gauging by implementing an air emissions monitoring network.
Timing the startup of the monitoring program with a plant expansion and a company desire to
focus on community safeguards, the community was kept well informed during public meetings
held during permit review. Kodak also planned to analyze methanol, acetone, ethanol, and
toluene releases.13
Web Site Provides Data for Pollution Prevention - Elyria, OH - 1998: After suffering from
poor health that appeared to improve when she was away from home, Pauline Leboda of Elyria,
Ohio began to suspect an environmental cause for her symptoms. She contacted Teresa Mills of
the Buckeye Environmental Network. Mills used the Scorecard web site, sponsored by the
C-5
-------
Environmental Defense Fund (www.scorecard.org), which provides TRI pollution data based
upon zip code, to discover what toxic emissions were causing the pungent odors in Elyna.
Further research proved that a local sponge manufacturer was operating without the proper toxic
emissions permit. In a settlement with the EPA, the sponge manufacturer agreed to pay a fine and
began using a chemical scrubber. Leboda and Mills proved that information can lead to action.14
Largest Environmental Polluter in State Attempts to Reduce Waste - Derry, NH - 1997:
After being identified as the largest environmental polluter in New Hampshire, senior management
at HADCO Corporation initiated efforts to reduce releases of toxic chemicals and transfers of
these substances from its facility. One of the nation's largest manufacturers of printed wiring
boards, HADCO used chlorinated solvents in their multi-step manufacturing process. To reduce
waste, HADCO implemented a solvent recovery system for some processes and eliminated the
use of chlorinated solvents for others, while installing a continuous emissions monitoring system
(CEMS). The company also switched to aqueous solvents, removing methylene chloride and
other toxics from its process.
As a result of these actions, HADCO eliminated the annual disposal of 800,000 pounds of
methylene chloride and no longer needed to operate the CEMS. Investment in this technology
paid for itself in three years.15
The Good Neighbor Project - Minnesota -1991: Following enactment of the 1990 Minnesota
Toxic Pollution Prevention Act, over 500 Minnesota manufacturers were required to compose a
pollution prevention plan and submit annual progress reports to the Minnesota Pollution Control
Agency. The Good Neighbor Project, initiated in 1991 by the Minnesota Citizens for a Better
Environment (CBE), encouraged community involvement in these pollution prevention plans.
The Good Neighbor Project also identified the toxic polluters with the greatest potential
environmental and health impacts and helped open a dialogue between the communities and
industry.
In January 1993, CBE released "Get to Know Your Local Polluter," a report that profiled the top
40 polluters in Minnesota and provided communities with applicable TRJ data, local
demographics on nearby sensitive populations, exposure scenarios, and potential effects of the
chemicals used or stored at each facility. The CBE is acting as the organizing body for many of
these Good Neighbor agreements. By using right-to-know and TRJ data, CEC is providing
communities with the information they need to play a role in local industry pollution prevention
and teaching industry how to work with the surrounding communities.16
Wake-up Call Reduces Pollution and Saves Money - Emigsville, York County, PA - 1994:
After Berg Electronics began reporting emissions under TRI, the company realized that they were
releasing almost 300,000 pounds of hazardous chemicals into the environment. By installing a
new cleaning system, the company reduced its emissions to 391 pounds per year. Although up-
front costs for the new system were relatively high ($500,000), the company now saves about
$1.2 million each year by avoiding cleanup and hazardous waste disposal costs.17
C-6
-------
TRI Data Used to Track Progress of the Big Three Auto Companies - Michigan - 1995: The
Ecology Center Toxics Reduction Project used TRI data to follow the pollution reduction
progress of the "big three" auto companies. Ecology Center was able to use TRI data to prove
flaws in emissions reduction analyses performed by these companies. In addition, the Center
worked with the Great Lakes Auto Pollution Prevention Alliance to initiate discussions between
plant management and local communities.
As a result of these discussions, the president of Auto Alliance International, which had been one
of the largest polluters by volume in the state, committed to an aggressive solvent reduction
program. The company's program will recapture solvents in the process, saving money and
improving air quality in the long run.18
Changing Users of TRI Data Reflect Benefits of EPCRA - 1990: When the TRI data first
became available in June 1988, industry topped the list of reviewers, primarily checking to see the
accuracy of their own data and how their releases compared to competitors. EPA began to notice
a change in the user trends at the close of 1989. More and more state agencies, environmental
officials, health care institutions, and citizens were beginning to use the TRI data in their attempts
to improve public health, local land use, and regulatory actions. The access to TRI data enabled
citizens and environmental groups alike to push for risk reduction, pollution prevention, and
stronger environmental laws.
For example, Louisiana used the TRI data to pass legislation aimed at cutting toxic air emissions
in half by 1994. Many states, such as Illinois and Indiana, now require property sellers to disclose
EPCRA-related information so that all buyers are aware of the past uses of the land. A South
Texas school district even used TRI data to locate a safe site for a new elementary school.
Around the country, citizens, environmental groups, and state and local authorities are working
with these data to create an accurate picture of the chemical risks in their community.19
TRI Yields Concrete Results, Researchers Find - U.S. - 1994: As a result of making TRI data
public and accessible, a majority of citizen groups and industry respondents surveyed in one study
reported that release reduction efforts were undertaken at plants, and that meetings were
prompted between industry and citizens. In addition, some facilities have signed "good neighbor
agreements," which include release reduction goals and citizen monitoring rights. The researchers
go on to state that "more voluntaristic approaches, built on forced leveling of the
information playing field, are a supplement to regulation for problems such as accident prevention
and toxics use reduction."20
Due to space considerations, some other successful uses of EPCRA data are referenced
below only briefly.
Public Pressure Leads to Toxic Use Reduction Laws - Massachusetts and Oregon - 1990:
TRI data used by public interest organizations, citizens, and legislators.21
C-7
-------
Colorado Manufacturer Takes First Steps, Makes Good Neighbor Pledge - Boulder
County, CO - 1991: Community and company action based on high TRI ranking.22
Citizens Win Funding for Refinery Monitoring - British Petroleum, Lima, OH - 1989: The
Facility was the biggest air polluter in the state, per TRI.23
Regulations Spurred by TRI - Louisiana - 1990: TRI data brought about public awareness,
which forced state policymakers to work towards a significant reduction of air toxics.24
TRI Data and GIS Used to Prioritize Pollution Prevention Effort - New Jersey - 1993:
New Jersey used TRI data with with geographic information systems (GIS) technology to map
the data and impacts.25
TRI Data Act as Catalyst, Address Environmental Injustice Issue - Ohio - 1995:
Environmental organization claims to show disproportionate waste and water impacts.26
Manufacturer Cleans Up After Several TRI Appearances as a Top NY Polluter - Lockport,
NY - 1996: Action came after appearing in TRI reports as one of state's top polluters.27
TRI Data Used to Change State Legislation - Utah - 1991: TRI data helped Sierra Club and
legislature identify problems.28
EPCRA Data Help Detail Correct Response to Midwest Floods - 1993: Local governments,
emergency responders, companies benefitted from their contingency plans built around EPCRA
data.29
TRI Data Lend Support to "Environmental Justice" Debate - Chicago, IL - 1991:
Environmental group used TRI data to call attention to hazardous waste transporters from out of
state.30
Facilitating Cooperation Through Committee - Tennessee - 1994: State that ranked high in
TRI emissions formed state, industry, and non-profit committees to analyze TRI data, reduce
pollution, and increase public awareness.31
TRI Data Help Identify Health Risks - New York - 1995: The State Department of Health
used TRI data to develop rankings that suggested health risks that could result from toxic
releases.32
Publicity of TRI Proves To Be Motivating Factor - Westwego, LA - 1992: American
Cyanamid launched program aimed at cutting TRI emissions by 80 percent.33
TRI Data Prove to be Powerful Tool in Stopping Polluters - Baton Rouge, LA - 1995: An
environmental organization uses TRI data to identify polluters and promote environmental
justice.34
C-8
-------
Community/Manufacturer Agreement Improves Plant Safety and Relations - Manchester,
TX - 199335
TRI Data for U.S. Companies Used to Assess Facilities in Mexico - Texas - 199436
Organized Community Defeats Ammonia Facility Application - Cloverleaf, TX37
B. USES OF RIGHT-TO-KNOW INFORMATION ASIDE FROM EPCRA
States with Right-to-Know Programs Reduce Emissions Significantly - U.S. - 1995/1996:
Several states have their own environmental right-to-know programs; studies of these can reveal
impacts likely in national programs. In one study, researchers used data from the TRI to conduct
a preliminary analysis of the effectiveness of right-to-know programs in reducing reported releases
across the 50 states. They found that states with functional right-to-know programs are
significantly more successful in reducing in-state toxic emissions that states without them. In
addition, they found that the effect of a right-to-know program on toxic releases outside a state
was not significant, contradicting an argument that such self-protection policies shift pollution to
other states.38
Toxics Use Down 20 Percent in Massachusetts - Boston, MA - 1997: In the four years
following enactment of the Toxics Use Reduction Act (TURA) of 1990 in Massachusetts, the use
of toxic chemicals by companies in the state dropped 20 percent and the volume of chemicals
ending up as waste also fell by 30 percent. While TURA does not mandate any process changes,
it does require facilities to disclose what chemicals they use and the waste that they generate. The
lack of any direct mandates for change in industry raised doubts among some companies and
industry associations. However, the state government and environmental organizations such as
MASSPIRG, argue that the public disclosure of TURA and TRI data seems to be the motivating
factor for the reduction in toxic chemical use as manufacturers look for cost-effective and safer
alternatives. Paul Burns of MASSPIRG noted that "the bottom line is that citizens have the right
to know when and how they are exposed to toxic chemicals. This law [TURA] has been a
powerful tool in getting businesses to voluntarily change their behavior."39
Indonesia Plants Clean Up to Make Better Pollution Grade - Indonesia - 1999: Indonesia's
Environmental Impact Management Agency ran a pilot program, known as "PROPER," in which
certain industrial facilities were graded, based on their water pollution performance. Researchers
found that, "if [reputation effects] are important, then market agents and communities, once
properly and accurately informed, can interact with firms to establish jointly-optimal levels of
consumption and production."40
Disclosure of the Indonesian facilities' grades was sufficient to prompt 10 factories to invest in
pollution abatement in order to improve their rating, and lead to a more than 40 percent pollution
reduction in the pilot group in only 18 months. Other countries are adopting similar programs. In
the Philippines, the national environmental agency's "EcoWatch" program has already used
C-9
-------
disclosure to dramatically increase compliance among 52 factories. It seems that if the U.S. were
to significantly restrict right-to-know, it would run counter to a broad and growing international
trend.41
Impact of Public Information in Canada: A 1999 study funded by the World Bank, Incentives
for Pollution Control: Regulation And (?) Or (?) Information, analyzed regulatory enforcement,
public information, and the relationship between the two (noted by the question marks in the title)
regarding emissions in British Columbia. The study found that clear, strong standards with a
significant and credible penalties produce emissions reductions. The authors also found that "the
public disclosure of environmental performance does create additional and strong incentives for
pollution control." They found evidence that the impact of public information was stronger than
that of fines. They concluded that the combination of regulations and information puts different
kinds of pressure on firms, "increasing the likelihood that they will undertake actions in line with
environmental protection.'"42
Appliance Labelling for Energy Consumption Encourages Environmental and Cost Savings
- U.S. - 1995: Ensuring that the public has access to information can be achieved in many ways.
One way is product labeling. For instance, large appliances now carry labels describing expected
energy usage and costs. It is well documented that, in the past, appliance buyers frequently did
not buy energy-efficient devices, despite the fact that the rate of return due to energy cost savings
far exceeded that of other investments. This phenomenon is attributed to multiple factors, one
being the transaction cost of transferring information known by the seller to a potential buyer.
Increasing the information available to consumers via standardized labels reduces this barrier, and
can both benefit the environment and save the consumer money.43
CONCLUSION
Granted, labeling products differs from disseminating information about community hazards via
paper or the Internet. However, the above examples show that improving the ease of acquiring
such information clearly contributes to behavioral change by individuals and companies. While
there are differences in the types of information and in the means of access in the programs above,
the trend is clear. Information that is readily available and relevant to the well-being of
individuals often leads to improvements for the interested parties - and often for the larger
public as well.
C-10
-------
Daniel
Campanelli/R3/USEPA/US
07/06/2005 08:34 AM
Hanrahan/R3/USEPA/US@EPA
bcc
To Hector Gerena/R3/USEPA/US@EPA
cc KarenD Johnson/R3/USEPA/US@EPA, Grace
Subject C Drive Transfer of Folders from Dan's Old PC
Hector,
The following folders (containing the files) from my C Drive will have to be transferred from my old
computer to my new computer.
C: My Files
C: DTFWRITR
C: DTFWRITR2
C: ACTIONSDTF/DTF
Please note that since there were problems with file transfer in the past, I definitely want to be present at
mv computer when the file transfer takes place.
Can you ask Dan to produce an itemized list identifying the folders and files that he wants to transfer to his
new computer? The itemized list is needed to expedite the SDWIS installation, configuration and file
transfer processes. CSB has lent me an EDI cable, used for file transfers, this week only. These cables
are constantly used by CSB and it will have to be returned as soon as I complete Dan's job. I would like
to finish the job this week, and if I could have a list from Dan by COB 7/6,1 could finish this work as I
propose. If I do not have the list to prepare myself, I am afraid that the file transfers will have to be done
by CSB at their convenience.
Regards,
...Dan...
Hector
Gerena/R3/USEPA/US
07/05/2005 07:43 PM
cc Jacqueline Pine/R3/USEPA/US@EPA, Grace
Hanrahan/R3/USEPA/US@EPA
Subject Request of Folder and File Names for PC Transfer
To KarenD Johnson/R3/USEPA/US@EPA
Karen -
Thank you
Forwarded by Daniel Campanelli/R3/USEPA/US on 07/06/2005 08:31 AM
Daniel
Campanelli/R3/USEPA/US
06/03/2005 10:32 AM
cc KarenD Johnson/R3/USEPA/US@EPA, Jacqueline
Pine/R3/USEPA/US@EPA
Subject Set-Up of Dan's New PC
To Hector Gerena/R3/USEPA/US
-------
Hector,
This is to confirm that you will be available to set up my new PC during the week of July 5th to July 8th,
2005. I will need your assistance because of the following:
- Transfer of latest C Drive files from the C Drive of my old PC to the C Drive of my new PC
- DTF Writer (SDWIS/Fed) Version 5.2 (DOS)
- DTF Writer (SDWIS/Fed) Version 5.5 (DOS)
- ActionsDTF (SDWIS/Fed) Versionl .2 (MS Access .mdb)
- ODD (SDWIS/Fed) Version 3.91
- ECDB (SDWIS/Fed) Version 3.91
- SDWIS/Fed Tracking System (Version ,02.mde or later)
- FedRep Version 1.2 (including TomCat, Java, etc.) and the related Patch for FedRep 1.2
- XML Writer (latest version)
- DynaComm (Series 8)
-FTP
- PWSS (latest version)
- UIC Tool (Arc View)
- WinZIP
- Lotus (Lotus Smart Suite)
- Word Perfect
- MS Access
- MS Word
- MS Excel
- MS Power Point
- Acrobat reader
- Adobe Reader (6.0 or later)
- Internet Explorer
- Netscape Communicator
- Lotus Notes (Version 6.5.2 or later)
- Printers setup
Thanks,
...Dan...
-------
hectorlgT.txt
Files
Size
Last update
Owner
= Files contained in this path
= Number of bytes in the file
= Date file was last updated
= ID of user who created or copied the file
CLSTRlSANUlB/OPM:GROUP\FINANCE\TRANSIT\*.*
Files Size Last update
Owner
TRANCURR.WK4 O
TRANCURR.XLS 0
TRB17T-1.WK4 0
417,300 6/23/05 11:17a Hgerena
445,952 2/02/05 7:45a Hgerena
402,812 4/18/03 12:33p Hgerena
1,266,064
3
bytes (1,269,760
Fi 1 es
bytes in 310 blocks allocated)
Page 1
-------
APPENDIX C REFERENCES
1. Tryens, Jeffrey, Richard Schrader, and Paul Orum. Making the Difference: Using the Right-
to-Know in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and
Working Group on Community Right-to-Know. (undated): p. 1. (Based on interview with Ted
Smith, Silicon Valley Toxic Coalition.)
2. U.S. EPA. An Overview of Uses of the Toxics Release Inventory Data in the U.S.. Susan B.
Hazen. Office of Pollution Prevention and Toxics. (June 1995): p. 11. (Based on Faye Rice, "Who
Scores Best on the Environment." Fortune, Vol.128, No.2. (July 26, 1993): p.l 14-122.)
3. Settina, Nina and Paul Orum. Making the Difference, Part II: More Uses of Right-to-Know in
the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and Working Group
on Community Right-to-Know. (October 1991): p. 11-12.
Contact: Stuart Greenberg, Environmental Health Watch
4. Tryens, Jeffrey, Richard Schrader, and Paul Orum. Making the Difference: Using (he Right-to-
Know in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and Working
Group on Community Right-to-Know. (undated): p. 15. (Based on article and interview with
Jane Nogaki, New Jersey Coalition Against Toxics).
5. Michuda, Colleen. TRI Success Stories: Citizens Making a Difference. U.S. EPA. (August
1994): p. 6.
Contact: Diane Hemingway, Director, CEC - Western NY.
6. Selcraig, Bruce. "What You Don't Know Can Hurt You." Sierra. Vol. 82, No.l. (Jan-Feb
1997):.p.38-45.
7.Tryens, Jeffrey, Richard Schrader, and Paul Orum. Making the Difference: Using the Right-to-
Know in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and Working
Group on Community Right-to-Know. (undated): p. 14. (Based on interview with Michael
Belliveau, Communities for a Better Environment).
8. Mundy, Dwayne. e-mail November 11, 1999.
9.Tryens, Jeffrey, Richard Schrader, and Paul Orum. Making the Difference: Using the Right-to-
Know in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and Working
Group on Community Right-to-Know. (undated): p. 12-13. (Based on interview with Ed Norman,
NC Environmental Defense Fund).
10. Heame, Shelley. "Tracking Toxics: Chemical Use and the Public's Right-to-Know."
Environment, Vol. 38, No. 6. (July-August 1996): p. 4-16.
11. Right-to-Know Planning Guide, "Emergency Planning..." Bureau of National Affairs, (May
C-ll
-------
1994): p. 2.
12. Settina, Nina and Paul Orum. Making the Difference, Part II: More Uses of Right-to-Know
in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and Working Group
on Community Right-to-Know. (October 1991): p. 3-4.
Contacts: Richard Metcalf, ACTWU, and Frank Wolf, Clean Air in Northfield.
13. BNA Reporter."Preventing Pollution..." Bureau ofNational Affairs (February 27, 1992): p. 3.
14. "EDF Letter." Environmental Defense Fund. December 1999, Vol. 30, No. 6, p. 5.
15. BNA Reporter. "Preventing Pollution..." Bureau ofNational Affairs, (December 25, 1997): p.
3. (discussing an EPA case study)
16. Michuda, Colleen. TRI Success Stories: Citizens Making a Difference. U.S. EPA. (August
1994): p. 10-11.
Contact: Jo Haberman, Citizens for a Better Environment - Minnesota.
17. Michuda, Colleen. TRI Success Stories: Citizens Making a Difference. U.S. EPA. (August
1994): p. 17. (Based on article in "Philadelphia Inquirer," June 18, 1994.)
18. The Right Stuff: Using the Toxics Release Inventory. OMB Watch and Unison Institute. (July
1995): p. 17.
19. Right-to-Know Planning Guide, "EPA Notices Change in Users of Toxic Release Inventory."
Bureau ofNational Affairs. (February 1990): p. 4.
20. Lynn, F. M. and J. Kartez. "Environmental Democracy in Action: The Toxics Release
Inventory," Environmental Management, Vol. 18, No. 4, pp. 511-521, 1994, pp 517-519.
21 .Tryens, Jeffrey, Richard Schrader, and Paul Orum. Making the Difference: Using the Right-
to-Know in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and
Working Group on Community Right-to-Know. (undated): p. 2. (Based on interview with Marc
Osten, MASSPIRG and Quincy Sugarman, OSPIRG.)
22. Settina, Nina and Paul Orum. Making the Difference, Part II: More Uses of Right-to-Know
in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and Working Group
on Community Right-to-Know. (October 1991): p. 7-8.
Contact: Larry Bulling, Colorado Citizen Action.
23.Tryens, Jeffrey, Richard Schrader, and Paul Orum. Making the Difference: Using the Right-to-
Know in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and Working
Group on Community Right-to-Know. (undated): p. 9. (Based on interview with Ed Hopkins,
Ohio Citizen Action).
C-12
-------
24. Tryens, Jeffrey, Richard Schrader, and Paul Orum. Making the Difference: Using the Right-
to-Know in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and
Working Group on Community Right-to-Kriow. (undated): p. 12-13. (Based on interviews with
Daryl Malek-Wiley, LA Sierra Club; Elouise Well, LA Environmental Action Network).
25. U.S. EPA. An Overview of Uses of the Toxics Release Inventory Data in the U.S.. Susan B.
Hazen.Office of Pollution Prevention and Toxics. (June 1995): p.9.
26. The Right Stuff: Using the Toxics Release Inventory. OMB Watch and Unison Institute. (July
1995): p. 34.
27. BNA Reporter. "Preventing Pollution..." Bureau of National Affairs. (May 2, 1996): p. 3.
28. The Right Stuff: Using the Toxics Release Inventory. OMB Watch and Unison Institute. (July
1995): p. 28.
29. Right-to-Know Planning Guide. "EPCRA Data Plays Major Role in Midwest Flood
Response." Bureau of National Affairs. (July 1993): p. 4.
30. The Right Stuff: Using the Toxics Release Inventory. OMB Watch and Unison Institute. (July
1995): p. 16.
31.Michuda, Colleen. TRI Success Stories: Citizens Making a Difference. U.S. EPA. (August
1994): p. 19. Contacts: Alan Jones, Tennessee Environmental Council and Angie Pitcock,
Division Director, TN Dept. of Environment and Conservation.
32. U.S. EPA. An Overview of Uses of the Toxics Release Inventory Data in the U.S.. Susan B.
Hazen. Office of Pollution Prevention and Toxics. (June 1995): p.7.
33. David Rotman. "Cleaning up processes," Chemical Week. Vol. 150, No. 23. (June 17,
1992).
34. The Right Stuff: Using the Toxics Release Inventory. OMB Watch and Unison Institute. (July
1995): p. 30-31.
35. Right-to-Know Planning Guide, "Partnership Linked to Improvements at Chemical Plant."
Bureau of National Affairs. (August 1993): p. 4.
36. The Right Stuff: Using the Toxics Release Inventory. OMB Watch and Unison Institute. (July
1995): p. 19.
37. Settina, Nina and Paul Orum. Making the Difference, Part II: More Uses of Right-to-Know
in the Fight Against Toxics. Washington, DC: Center For Policy Alternatives and Working Group
on Community Right-to-Know. (October 1991): p. 15-16.
C-13
-------
Contacts: Karla Land, North Channel Concerned Citizens Against Pollution; Jim Baldauf, Texans
United; and Dr. Fred Millar, Friends of the Earth.
38. Grant, D.S. and Liam Downey. "Regulation though Information: An Empirical Analysis of
the Effects of State-sponsored Right-to-know Programs on Industrial Toxic Pollution." Policy
Studies Review, (Autumn/Winter 1995/1996): p 346.
39. Right-to-Know Planning Guide. "Massachusetts Reports 20 Percent Drop in Toxics Use."
Bureau of National Affairs. (January 1997): p. 4.
40. Afsah, Shake, Benoit Laplante, and David Wheeler. "Regulation in the Information Age:
Indonesian Public Information Program for Environmental Management." (March 1997): p. 5, 8.
http://www.worldbank.org/nipr/work_paper/govern/index.htm.
41. World Bank, "Greening Industry: New Roles for Communities, Markets, and Governments."
Oxford University Press. (1999): pp 64-74.
42. Foulon, J., P. Lanoie, and B. Laplante. "Incentives for Pollution Control: Regulation And (?)
Or (?) Information," (October 1999)
http://www. worldbank.org/nipr/work_paper/andor/index. htm.
43. Howarth and Sanstad, "Discount rates and energy efficiency," Contemporary Economic
Policy. Vol. 13, No. 3. (July 1995): p. 101(9).
C-14
-------
This page intentionally left blank.
C-15
-------
APPENDIX D
HOW DOES NEGATIVE PRESS AFFECT
FACILITIES' TOXIC RELEASE EMISSIONS?
D-l
-------
APPENDIX D
HOW DOES NEGATIVE PRESS AFFECT
FACILITIES' TOXIC RELEASE EMISSIONS:
AN INFORMAL ANALYSIS
Introduction
The Toxic Release Inventory (TRI) database, which requires manufacturing facilities that
release to the environment any of 300 chemicals to file an emissions report with EPA, is a major
component of the Emergency Planning and Community Right-to-Know Act (EPCRA). Every
citizen can access this database and can determine the presence and release of hazardous and toxic
chemicals at industrial locations. In this analysis, we wanted to discover whether and to what
extent negative press coverage might affect a facility's TRI emissions. Does publicly provided
information on toxic chemical releases lead to emissions reductions? When newspapers target a
facility for its large amount of toxic emissions, does that facility reduce its toxic releases in
subsequent years? Does newspaper and media criticism change corporate behavior and the
allocation of corporate resources to lessen the negative criticism? This Appendix describes the
results of an informal, case-based analysis on whether media criticism appears to have any positive
effect on toxic emissions. It is not intended to represent a robust statistical analysis of this
hypothesis.
Approach
Several different searches for newspaper, magazine, or trade journal articles that cited
Toxic Release Inventory data on one or more specific facilities were performed. The search
criteria included keywords such as routine or annual toxic emission, toxic releases, TRI or Toxic
Release Inventory or Section 313 or EPCRA or Community Right-To-Know, etc. Although a
number of facilities were mentioned in various articles, seven facilities seem to have been cited
most often in article after article, year after year, as "the worst polluters" in the country,
according to their TRI emissions. In addition, a number of facilities were identified that were
cited as "the worst polluters" in their states, according to their TRI emissions. The years in which
these facilities were targeted by the media for their high level of emissions were also noted. The
newspaper articles identified these facilities for the sheer quantity of their TRI emissions. For
example, one facility could be responsible for 80 percent of a state's total TRI emissions in a one-
year period.
The TRI database was used to compare the quantity of toxic release emissions for each of
the seven "worst polluters" before it received negative press to the quantity of toxic release
emissions after it received negative press. Obviously, there are many reasons for a company to
reduce its emissions. Companies with large emissions may be able to find ways to reduce those
emissions more quickly and easily than others and sometimes facility modifications to reduce
D-2
-------
emissions could take more than a year; some of these emission reductions may have been in the
planning stages prior to negative press coverage. Consequently, emission results for several years
after press coverage were included to see whether reductions in TRI emissions occurred sometime
after a facility received negative press.
To further judge the effect of negative press, we identified a comparable facility located in
the same region with the same Standard Industry Classification (SIC) code emitting the same
chemicals as each of the "worst polluter" facilities in particular states. However, these
comparable facilities did not receive negative press for their TRI emissions. For these comparable
facilities, we used the TRI database to track the quantity of toxic release emissions over the same
time period as the "worst polluting" facilities. We then compared the direction and rate of change
in TRI emissions for publicized facilities as compared to non-publicized facilities. As above,
emission reduction efforts can take time and can become more challenging as emissions are
reduced; a facility may have been planning emission reductions regardless of press coverage.
Results
Nationwide
A significant number of companies have made a wide variety of reductions in their TRI
emissions. However, as noted above, seven companies were selected for closer examination
because they repeatedly appeared in articles year after year because of the large quantities of toxic
chemicals they emit. As early as 1989, Du Pont, Monsanto, American Cyananiid, Kennecott, and
IMC-Agrico were targeted as the worst polluters in the country based on their total emissions in
1987. These companies appeared in newspapers all over the country, from the Los Angeles
Times, to the St. Louis Post Dispatch, to the Lexington-Herald Leader. A public interest group,
Citizen Action, named Inland Steel a top ten toxic polluter in 1991 and the Minneapolis-St. Paul
Business Journal named 3M among the nation's top polluters in 1991.
For a core group of chemicals, TRI emissions for nearly all companies have gone down
43% between 1988 and 1997. The core group of chemicals have remained the same since the first
TRI reports in 1987. In contrast, the companies which were publicly criticized for large quantities
of TRI emissions significantly reduced their emissions after receiving negative press. Those
companies named as some of the "worst polluters in the country" reduced their emissions as much
as twice the general TRI trend.
Between 1990 and 1996, the following companies significantly reduced their total toxic
releases and transfers. These reductions are much greater than the general trend in TRI emissions
reductions between 1991 and 1996.
D-3
-------
Table D-l - Reductions of TRI Emissions for Selected Companies
General
Reduction over
Increased Improvement
Percent
Trend
General Trend
over General Trend
Company
Reduction
(%)
(%)
(factor)
Inland Steel
95
43
52
2.2 times
Kennecott
90
43
47
2.1
Monsanto
84
43
41
2.0
American Cyanamid
83
43
40
1.9
IMC-Agrico
82
43
39
1.9
Du Pont
73
43
30
1.7
3M
65
43
22
1.5
Although a facility may reduce its toxic emissions for a number of reasons, "manufacturers
listed among the worst polluters... may change their ways out of fear of customer boycotts,
increased regulation, or community hostility. The company's reputation, hard to build and easy to
destroy, is at stake" (D55).
Ohio
Between 1989 and 1992, Ohio newspapers named Honda of America and O.M. Scott and
Sons the highest polluters in the state. In 1989, the Columbus Dispatch reported that O.M. Scott
was top polluter because of its ammonia releases. In 1991, the Cleveland Plain Dealer reported
on Honda's toxic releases, the highest in the state. The Columbus Business First criticized both
Honda and O.M. Scott for their toxic releases in a 1992 article. In addition, Ohio EPA cited
Honda of America for nine air pollution violations in 1992. Between 1991 and 1994, Honda
reduced its total toxic releases and transfers by 30 percent. However, since 1994, total on-site
emissions by Honda have continued to decrease although off-site transfers have increased beyond
the 1991 levels. O.M. Scott and Sons (now Scotts Company) reduced its total toxic releases and
transfers 52 percent between 1990 and 1994. O.M. Scott's emission levels rose again in 1995 to
almost the 1991 level, but declined again in 1996 and 1997.
Tennessee
In 1992, USA Today named the top 15 "toxic offenders" that release the most toxic
material (in pounds). Four of Du Pont's facilities [New Johnsonville, Tennessee (6,h of the 15),
Beaumont, Texas (11th), Pass Christian, Mississippi (12th), and Victoria, Texas (14th)] were
included in the list. Long named the nation's top polluter, Du Pont is cited in the press more than
any other company for its toxic emissions. However, between 1991 and 1996, Du Pont reduced
its emissions by 73% company-wide.
In 1991, the Memphis Commercial Appeal reported on Du Pont's pollution-control
programs. In 1992, the Commercial Appeal reported that Tennessee was the nation's third
largest producer of toxic waste in 1990, with Du Pont among the state's top 10 dischargers. But,
D-4
-------
in 1994, the Memphis Business Journal openly criticized Du Pont's New Johnsonville plant for
the largest emissions of toxic chemicals in the state. Du Pont's New Johnsonville plant reduced
its total toxic releases and transfers by 97 percent between 1990 and 1996; 94 percent of this was
between 1994 and 1995. In 1991, the New Johnsonville plant accounted for 19% of Tennessee's
total toxic releases. In 1996 however, this same plant accounted for 1 percent of Tennessee's
total toxic releases. Certainly, negative press coverage appears to have helped spur corporate
spending and improved behavior. Note that a company may need time to enact modifications
between when negative press is published and changes in the amount of emissions.
Du Pont, New Johnsonville. TN
70.000.000
60.000.000
=- 50.000.000
»
o
M
2 40.000.000
C
2 30.000.000
o
V)
re
£ 20.000.000
-------
Kemlra Pigments, Inc , Savannah. GA
SIC 2816 (Ou Pont, TN)
>.200.000
.000.000
000,000
600.000
400.000
0
1004
1065
1996
1 000
1
Montana
Asarco Inc.'s East Helena plant has been publicized as the top emitter of toxic releases in
Montana since 1990. The American Metal Market reported in September 1992, that Asarco
"ranked seventh on the list of manufacturing companies with the most toxic chemical releases."
The company defended its ranking by explaining that "eighty-seven percent of its releases are
made up of slag, which when properly managed presents no hazard to the public. " The Billings
Gazette reported that Asarco's East Helena plant led the 1993 list of Montana companies in
discharges into the air, water and land in the course of manufacturing. Between 1990 and 1996,
the East Helena plant increased its toxic emissions by 12 percent, while company-wide, Asarco
Inc. increased its total toxic emissions by 50%. On average, Asarco, Inc. East Helena accounts
for 92 percent of Montana's total toxic emissions between 1990 and 1996.
The Exxon Billings Refinery has also been in the Montana press for its toxic emissions.
While The Billings Gazette named Asarco for the quantity of toxic chemicals it released, Exxon
was named in 1991 for its release of xylene. Xylene can cause birth defects, and at high levels can
cause dizziness, passing out and even death. Between 1990 and 1996, Exxon reduced its total
xylene releases and transfers by 86 percent. Over the same time period, the Exxon Billings
Refinery reduced its total releases and transfers by 64 percent. Whereas, for all of Exxon's
facilities, the total toxic releases and transfers has stayed the same between 1990 and 1996.
Company-wide, Exxon has not improved their total toxic emissions. But, in Montana, where
Exxon received negative press, the Billings Refinery reduced their total TRJ emissions.
D-6
-------
Exxon Billings Refinery, Billings, MT
700,000
600.000
500.000
400,000
5 300.000
200.000
100.000
960
1992
1994
1
991
1
Louisiana
As early as 1987, Louisiana newspapers named a portion of the state "Cancer Alley." 15
chemical companies reside along a stretch of the Mississippi River near St. Gabriel. Residents
blamed the pollution from these facilities for the area's "alleged high miscarriage, stillbirth, and
cancer rates" according to a 1987 article in Industiy Week. Since the first TRI data became
available to the public in 1989, Louisiana has had the highest toxic emissions in the country. Four
facilities in "Cancer Alley," American Cyanamid Fortier Plant, Agrico Chemical Faustina, Agrico
Chemical Uncle Sam, and Arcadian Fertilizer, account for approximately 60 percent of
Louisiana's total toxic releases. Not only was Louisiana named the worst polluting state in the
country, but these four facilities were frequently in the Louisiana newspapers between 1989 and
1994.
In 1990, The New Orleans Times Picayune had two articles on the top polluters in
Louisiana. One article states, "American Cyanamid's [facility in Waggaman] releases are greater
than those of the entire state of New Jersey. Since New Jersey is 15"' (in the nation) in total
emission, this means American Cyanamid's releases are greater than the total releases of most
states." In 1991, four articles in the New Orleans Times Picayune reported on Agrico's and
Arcadian's high toxic emissions. Also in 1991, the Picayune called Louisiana "the Wasteland."
They reported that American Cyanamid "discharged more toxic waste underground than the total
discharged by any one of the 48 states in 1988." In 1992, the Picayune again criticized American
Cyanamid for its "dubious distinction of being the most polluting plant in the country." Finally, in
1993 and 1994, the Louisiana Industiy Environmental Alert reported on the total TRI emissions
of Agrico, Arcadian, and American Cyanamid.
D-7
-------
The total toxic emissions for Agrico Faustina and Agrico Uncle Sam steadily increased
between 1990 and 1993. Arcadian Fertilizer increased its toxic emissions between 1990 and
1992. All of these companies drastically reduced their emissions by 1994. Between 1990 and
1994, American Cyanamid reduced its total toxic releases and transfers by 87%. All four of these
facilities have kept their total TR1 emissions at the same levels since 1994.
Table D-2 - TRI Reductions for Selected Facilities in Louisiana
Facility
Percent Increase
(pre-1994)
Percent Reduction
(by 1994)
Agrico Faustina
126
90
Agrico Uncle Sam
71
94
Arcadian Fertilizer
97
68
LOUISIANA, Total Releases and Transfers
56
67
IMC-Agrico Faustina Plant, Saint James, Louisiana
140,000,000
120.000.000
J 100,000,000
80.000,000
60,000.000
40.000.000
20,000.000
1990
1991
1992
1993
1994
1995
1996
Year
D-8
-------
IMC-Agrico, Uncle Sam, Lousiana
70.000.000
60.000.000
50.000,000
40.000.000
30.000,000
20.000.000
10.000.000
1992
1993
1991
1994
996
1995
1
Year
Arcadian Fertilizer, Louisiana
45,000.000
40.000,000 •
35,000,000
-------
Cytec IncL Inc., Westwego, LA
(American Cyanamid)
180,000,000
160,000,000
140,000,000
I
^ 120.000,000
£
(ft
C
c 100,000,000
¦o
C
«o
S 80,000,000
<0
(0
o
®
? 60,000,000
<0
o
H
40,000,000
20,000,000
0
For the two Agrico facilities, we identified comparable facilities with which to compare
their total toxic releases and transfers. None of these comparable facilities showed such
significant reductions in toxic releases and transfers as those identified "worst polluters." Also,
these comparable facilities were not publicly criticized.
We compared Agrico Faustina to Air Products and Chemicals, Inc. Total toxic releases
for Air Products and Chemicals, Inc. declined slightly 1991 and 1994. However, their 1995 toxic
release levels matched the 1991 levels. We compared Agrico Uncle Sam to Farmland Hydro L.P.
The total toxic releases and transfers for Farmland Hydro increased from 1.47 million pounds in
1991 to 2.37 million pounds in 1996, a 61% increase.
Conclusions
• Many companies that received negative press about their total toxic releases reduced their
emissions 1.5 to 2 times more than the general TRI trend in toxic releases.
• As seen above, in Tennessee, Du Pont dramatically reduced its total toxic releases after
receiving negative press. Company-wide, Du Pont has also reduced its total toxic
D-10
-------
emissions. The total releases of a facility comparable to Du Pont's New Johnsonville plant
(but which did not receive negative press), generally did not change.
In Montana, Asarco's East Helena plant increased their emissions after they received
negative press. However, the emissions increases at the Montana plant were slight (12
percent) compared to Asarco's company-wide total emissions (50 percent). The Exxon
Billings Refinery was named in the press for the quantity of xylene it released into the
environment. After this publicity, the refinery reduced the release of xylene by 86 percent
and overall emissions by 64 percent.
In Louisiana, the top four polluting facilities were frequently cited by the press between
1989 and 1993. By 1994, all of these facilities had significantly reduced their total toxic
releases. Also, when we compared facilities, those that did not receive negative publicity
did not significantly reduce their total toxic releases.
Nearly all individual facilities studied reduced their TRi emissions well beyond the national
trend in years following negative press coverage about their toxic emissions. In some
instances, negative press coverage regarding many facilities within a company preceded
significant reductions in that company's overall TRI emissions.
D-ll
-------
APPENDIX D REFERENCES
Dl. Abrahamson, Peggy. "Group Cites PD for Pollution." American Metal Market. The Gale
Group. DIALOG File 148. July 17, 1991.
D2. "America's Worst Toxic Polluters; 8 Companies with Poor Environmental Records. "
Business and Society Review. The Gale Group. DIALOG File 148. Winter, 1993.
D3. Anderson, Ed. "Plant's Emissions Outrank Entire States." New Orleans Times Picayune.
DIALOG File 706. December 11, 1990.
D4. . "Six Area Plants Among Louisiana's Top Polluters." New Orleans Times
Picayune. DIALOG File 706. April 28, 1990.
D5. Associated Press. "Group Releases 'Toxic 500' List: 24 Plants Produce a Third of the
Hazardous Waste, Statistics Show." The Orlando Sentinel. DIALOG File 705. August 11,
1989.
D6. Bertelson, Christine and Ahmed, Safir. "Clean Air: States Vary Approach. " St. Louis Post
Dispatch. DIALOG File 494. January 21, 1990.
D7. Brookes, Warren. "Pollution Curb, Learning Curve." The Washington Times. DIALOG
File 717. January 9, 1991.
D8. Brown, T.C. and Tatge, Mark. "Honda's Home No. 1 in Ohio for Toxic Air." The
Cleveland Plain Dealer. DIALOG File 725. December 19, 1991.
D9. Bukro, Casey. "Council Criticizes 8 Firms' Records on Environment." The Chicago
Tribune. DIALOG File 632. December 15, 1992.
D10. Charlier, Tom. "Big Polluters in Shelby Cutting Back Toxic Flow." The Memphis
Commercial Appeal. DIALOG File 740. July 22, 1991.
Dll. . "Tennessee is Third in Toxic Waste." The Memphis Commercial Appeal.
DIALOG File 740. May 28, 1992.
D12. Dias, Monica. "Report Cites Area Polluters State's Worst Include 2 Local Companies."
The Kentucky Post. DIALOG File 722. July 23, 1991.
D13. Downing, Bob. "Report Details Release of Toxic Chemicals in Ohio for Year. " Akron
Beacon Journal. Resp. Database Services. DIALOG File 9. March 26, 1995.
D-12
-------
D14. "Du Pont Reports 6% Rise in Toxic Chemical Releases." The Wall Street Journal. May
18, 1993.
D15. "Du Pont Reports Decline in 1990 Toxics Releases." Chemical Marketing Reporter. The
Gale Group. DIALOG File 16. July 8, 1991.
D16. Edwards, Randall. "Union County Gets Bad Rap." The Columbus Dispatch. DIALOG
File 495. June 14, 1989.
D17. Epstein, Robin. "Ranking the Heavy Emitters: Toxic Corporations." The Nation. The
Gale Group. DIALOG File 88. December 5, 1994.
D18. "FOE Report Claims Du Pont Worst Polluter." Chemical Marketing Reporter. The Gale
Group. DIALOG File 16. September 2, 1991.
D19. Fried, John J. "Who's Who in Toxic Emissions." The Philadelphia Inquirer. DIALOG
File 633. August 28, 1992.
D20. "Industrial Wastes: State-By-State Look." USA Today. DIALOG File 703. May 28,
1992.
D21. Johnson, Clair. "Ranking of Polluting Firms Worsens." The Billings Gazette. Bell &
Howell. DIALOG File 635. No date.
D22. Journal of Commerce. "A Few Factories Release Most of Toxic Pollutants, Study Says."
Sacramento Bee. DIALOG File 496. October 6, 1990.
D23. . "Bulk ofToxic Pollution Tied to Handful of Firms: EPA Finds 36% of Emissions
Spewed by 50 Plants." Lexington Herald-Leader. DIALOG File 721. October 6, 1990.
D24. Kurschner, Dale. "3M's Pollution Record: Cloudy and Clear." Minneapolis-St. Paul City
Business. The Gale Group. DIALOG File 148. December 9, 1991.
D25. Lambrecht, Bill. "Polluters Ranked." St. Louis Post Dispatch. DIALOG File 494. July
12, 1991.
D26. Lombardi, Bill. "39.1 Million Pounds of Toxic Chemicals Released." The Billings Gazette.
Bell & Howell. DIALOG File 635. No date.
D27. Los Angeles Times. "Toxic Chemical Releases Down by 9%." St. Paul Pioneer Press
Dispatch. DIALOG File 701. October 4, 1990.
D28. . "Toxic Waste Discharges are Reduced." St. Paul Pioneer Press Dispatch.
DIALOG File 701. October 4, 1990.
D-13
-------
D29. Lucas, Allison. "Du Pont: Fighting its Way to a Cleaner Standard: But Pollution Numbers
are Slow to Come Down." Chemical Week. The Gale Group. DIALOG File 148. June 1, 1994.
D30. Martz, Michael. "Prevention Seen as Profitable: Du Pont Finds Resource in Recycled
Solvent." The Richmond News Leader. Richmond Times Dispatch. DIALOG File 709. July 26,
1990.
D31. "Metals Rated for Pollution and for Improvement." American Metal Market. The Gale
Group. DIALOG File 16. April 26, 1994.
D32. "Monsanto, Olin and Du Pont Claim Sharp Emissions Cuts." Chemical Marketing
Reporter. The Gale Group. DIALOG File 16. July 15, 1991.
D33. Moos, Shawna. "Pollution-Prevention Power to the People." Technology Review. Bell &
Howell. DIALOG File 15. October, 1992.
D34. Nelson-Horchler, Joani. "Superfund Poses Major Problems for Industry." Industry Week.
The Gale Group. DIALOG File 148. September 21, 1987.
D35. O'Byme, James and Schleifstein, Mark. "In Harm's Way - Geismar/St. Gabriel By the
Numbers." New Orleans Times Picayune. DIALOG File 706. February 18, 1991.
D36. . "Louisiana Industries Nearly Halved Pollution, DEQ Report Says." New Orleans
Times Picayune. DIALOG File 706. October 4, 1991.
D37. . "Louisiana Leads U.S. In Toxins." New Orleans Times Picayune. DIALOG File
706. May 28, 1992.
D38 . "Louisiana's Top Ten Toxic Polluters. " New Orleans Times Picayune. DIALOG
File 706. February 18, 1991.
D39. . "Troubled Plant: Neighbors Would Like to Breathe Easier." New Orleans Times
Picayune. DIALOG File 706. February 18, 1991.
D40. . "Underground Hazards: Drinking Water Fears Spread With Wastes." New
Orleans Times Picayune. DIALOG File 706. March 25, 1991.
D41. Phillips, Michael M. "The Price of Industry: Tons of Pollutants 10 Million Pounds
Released in 1987, EPA Records Show." The Philadelphia Inquirer. DIALOG File 633. April
26, 1989.
D42. Pope, Charles. "State Moves Up the List as Polluter: EPA Ranks State 23rd in Toxic
Emissions." The State. DIALOG File 720. May 17, 1991.
D-14
-------
D43. "Responsible Care: American Cyanamid." Chemical Week. The Gale Group. DIALOG
File 16. June 17, 1992.
D44. Rice, Faye. "Who Scores Best on the Environment." Fortune. Bell & Howell. DIALOG
File 15. July 26, 1993.
D45. Roberts, Timothy. "Tennessee Firms Work to Shed State's'Dirty' Image." Memphis
Business Journal. The Gale Group. DIALOG File 148. July 11, 1994.
D46. Rutherford, Glenn. "Kentucky 21s1 in Pollution , Group Says Jefferson, Marshall, Daviess
Counties Emis Most Toxins in State, Figures Show." Lexington Herald-Leader. DIALOG File
721. July 24, 1991.
D47. Spevacek, Jennifer. "Toxic 500 Behind Three-Fourths of Pollution." The Washington
Times. DIALOG File 717. August 11, 1989.
D48. "TheTRI: Two Views." Louisiana Industry Environmental Alert. The Gale Group.
DIALOG File 16. May, 1993.
D49. "Thinking Small About Pollution; A 1992 Overview of Waste Minimization." Chemical
Business. The Gale Group. DIALOG File 16. June, 1992.
D50. "Three Suppliers Top Citizen Group's Polluter List." Rubber and Plastics News. The Gale
Group. DIALOG File 16. July 22, 1991.
D51. "Top Ozone-Depleting Emissions, State by State." USA Today. DIALOG File 703.
January 17, 1990.
D52. "TRI, TRI Again." Louisiana Industry Environmental Advisor. The Gale Group.
DIALOG File 16. March, 1994.
D53. Walters, Rebecca. "Rural County Climbs to No. 4 on State's Toxic List." Business First-
Columbus. The Gale Group. DIALOG File 148. April 13, 1992.
D54. Weir, David and Yamin, Priscilla. "Toxic Ten: America's Truant Corporations." Mother
Jones. The Gale Group. DIALOG File 88. Jan. - Feb., 1993.
D55. Graham, Mary; Regulation by Shaming; The Atlantic Monthly, April 2000.
D-15
-------
APPENDIX E
ANALYSIS OF ACCIDENT DATA
E-l
-------
APPENDIX E
ANALYSIS OF ACCIDENT DATA
RiMP Data
Facilities subject to the RMP rule are required to submit, as part of their RMP, information on all
serious accidents that occurred in the five years prior to the date of RMP submission. Because
RMPs were submitted from March through the end of 1999, both 1994 and 1999 represent partial
year data. The table below shows the industry sectors that reported accidents in their RMPs; these
sectors are chemical manufacturers (NAICS 325); petroleum refineries (NAICS 32411); facilities
that use ammonia for cold storage (food processors, food distributors, refrigerated warehouses,
and food warehouses, NAICS 311, 4224, 49312, 49313); pulp and paper mills (NAICS 3221);
chemical wholesalers (NAICS 42269); drinking water treatment plants (NAICS 22131); and
wastewater treatment plants (NAICS 22132).
Table E-l - RMP Accidents Reported by Industry Sector
Sector-
1994
1995
1996
1997
1998
1999
5 Year
TOTAL
Annual
Average*
Chemical
Manufacturers
63
104
128
145
126
52
618
126
Refineries
17
31
50
34
38
23
193
39
Cold Storage
30
64
80
92
96
29
391
78
Pulp and
Paper
6
28
24
20
22
7
107
21
Chemical
Wholesaler
5
11
16
27
22
6
87
17
Water
Treatment
11
24
18
20
29
14
116
23
POTWs**
12
19
22
24
24
9
110
22
* 5 year Total divided by 5; assumes part year data for '94 and '99 are together equivalent to a full year.
** Publicly Owned Treatment Works for waste-water treatment
ERNS Data
The Emergency Response Notification System (ERNS) is a database used to store information on
notifications of oil discharges and hazardous substances releases. The ERNS program is a
cooperative data sharing effort among the Environmental Protection Agency, the Department of
Transportation, and the National Response Center (NRC). ERNS provides the most
E-2
-------
comprehensive data compiled on notifications of oil discbarges and hazardous substance releases
in the United States. Since its inception in 1986, more than 275,000 release notifications have
been entered into ERNS.
ERNS releases for four states — MA, CT, NJ, and VA — were reviewed. The states were
selected because of their similarity in size and industrial sectors. Releases were divided into three
categories:
• Fixed facility hazardous substance releases. Releases from unknown sources, those
generated by private citizens, and releases of unknown materials were deleted. Releases
of building products, such as asbestos and asphalt, were also eliminated.
Transportation releases, including all releases classified as transportation even if the source
or substance was unknown.
• All releases of oil and oil products. Non-petroleum oil releases were deleted (e.g., mineral
oil); releases listed as unknown oil were included.
Unless they were classified as transportation releases, a number of releases reported to ERNS
were eliminated and do not appear in any of the analyses. These releases were of unknown
substances (either listed as unknown material or unknown chemical or by vague descriptions, such
as "smoky debacle," "green glowing stick," "chemical odor," and "smells like cat urine") or of
substances clearly not subject to reporting (household appliances, used cars, injured duck, and
tires). Many releases appear in both transportation and oil releases (petroleum products may
represent up to 75 percent of transportation releases).
Hazardous substance releases were further analyzed as follows:
• Releases of currently listed TRI chemicals from facilities that appear to be manufacturers.
Where the type of business was unclear, but could be a manufacturer, it was included.
Releases of currently listed hazardous substances (including those listed by category)
where the reported quantity is greater than the current reportable quantity.
Many of the reported releases of TRI chemicals are not reportable under CERCLA; this is
particularly hue of the CFCs (e.g., freon) and CFC substitutes.
E-3
-------
Table E-2 - All Hazardous Substance Releases
Year:
CT
MA
VA
NJ
90
61
94
104
202
91
70
106
116
175
92
109
145
118
124
93
124
150
134
80
94
113
128
127
102
95
48
73
83
75
96
52
55
76
79
97
25
55
49
59
\o
CO
25
58
73
49
99
.21
44
85
49
Table E-3 - Releases of Current TRI Chemicals from Manufacturers
Year:
CT
MA
VA
NJ
90
38
45
36
151
91
43
43
46
82
92
57
39
42
61
93
56
51
49
44
94
63
42
55
44
95
26
32
24
28
96
13
21
33
30
97
12
15
22
32
98
9
16
24
22
99
11
17
29
28
E-4
-------
Table E-4 - Releases of Hazardous Substances Above the Current Reportable Quantity
Year:
MA
CT
VA
NJ
90
23
29
58
100
91
30
42
66
53
92
34
56
32
53
93
27
29
63
25
94
23
30
41
33
95
21
21
35
21
96
13
14
26
22
97
13
7
17
27
98
5
10
29
20
99
7
6
41
13
Table E-5 - Total Hazardous Substance Releases for Four States
Year:
All Hazardous
Substance Releases
TR1 Manufacturer
Releases
Reportable
Releases
90
461
270
210
91
467
214
191
92
496
199
175
93
488
200
144
94
470
204
127
95
279
110
98
96
262
97
75
97
00
CO
81
64
98
205
71
64
99
199
85
67
1999/peak
40%
31%
32%
E-5
-------
Table E-6 - Oil Releases
Year:
MA
CT
VA
NJ
90
612
113
535
503
91
700
128
510
518
92
701
151
806
483
93
616
184
822
479
94
669
281
876
645
95
364
194
743
629
96
539
359
717
659
97
447
246
527
418
98
473
271
794
415
99
368
260
754
442
Table E-7 - Transportation Releases
Year:
MA
CT
VA
NJ
90
165
58
285
295
91
182
52
384
293
92
208
52
295
253
93
192
81
434
231
94
215
117
493
312
95
115
86
405
297
96
136
115
450
314
97
110
93
355
189
98
195
87
465
214
99
104
113
450
239
E-6
-------
APPENDIX F
INSTITUTIONAL USES OF OCA DATA
CREATING INCENTIVES FOR RISK REDUCTION
F-l
-------
APPENDIX F
INSTITUTIONAL USES OF OCA DATA
CREATING INCENTIVES FOR RISK REDUCTION
The power of public scrutiny can manifest itself in many ways. Research reveals that
many types of institutions and individuals have interests that can be served by accessing Off-site
Consequence Analysis (OCA) information, as noted in Chapter 3. This appendix examines
segments of the public, their interests, their needs for access to OCA information, and their likely
uses of it. The nature of the access desired is also discussed.
Segments of the public discussed include:
~ Residents and community, public interest, and environmental organizations
~ News media
~ Emergency planning and response organizations
~ Industry and trade associations
~ Those with a financial interest in the company
~ Local officials and major local financial stakeholders
~ Workers and labor unions
~ Universities and research organizations
~ Professional organizations
Please note the following. An interested party may act with risk reduction as an explicit
goal, or it may act out of other interests (e.g., a newspaper seeking to improve its circulation) and
still lead to risk reduction. Also, except where noted, the parties' access described here is as
members of the public, rather than as covered persons or qualified researchers under the Chemical
Safety Information, Site Security and Fuels Regulatory Relief Act (CSISSFRRA).
Residents and Community, Public Interest, and Environmental Organizations
Residents and non-governmental organizations have a keen interest in access to
OCA information. The greater the disclosure, the more likely it is that citizens
will engage in dialogue with companies, governments, and their families to
address chemical hazards. In addition, there is ample evidence that for
information to be used, it must be easily accessed. Local residents and local
groups are most interested in information about facilities nearby. More
broadly-based organizations need the ability to compare the full range of
facilities across the nation.
F-2
-------
In 1992, a non-governmental organization in West Virginia's Kanawha Valley, along with
the National Institute for Chemical Studies, petitioned local chemical facilities to disclose worst
case accident scenarios. Two years later area chemical companies voluntarily did so at a public
event held where the most residents would normally be found - in a shopping mall. Although
there were concerns about panic or strong negative reaction in the community, these did not
occur. Instead, the effort led to a number of positive outcomes.1 Unfortunately, although the
Kanawha Valley experience was positive, such voluntary disclosures are rare, and generally will
not serve the needs of the vast majority of communities.
Wide disclosure of OCA information would inform citizens and citizens' groups, which
generally do not have access as covered persons under the statute. If the information raised their
concerns, it would stimulate activity. For example, they would more likely register concerns with
company management and request risk-reducing steps be taken. Likewise, they would more likely
contact local and state officials to register concerns and request that the officials take steps - from
requesting voluntary action of a facility to changing its operating permit. On an individual and
family basis, informed residents can take self-protective actions, such as learning about emergency
response and sheltering-in-place, participating in emergency drills, or moving residences or
children's daycare and school locations. Note that many of these actions can reduce the risk
posed by a chemical release, no matterwhether accidental or not.
In addition, concerned citizens may also seek information about facilities in other counties
or states to compare their situation with that of friends, relatives, or others. Real estate guides
now recommend that families find out environmental information about particular neighborhoods
on the Internet before they relocate.2 Ideally, citizens and organizers could examine the OCA
information (both worst-case and alternative release scenarios), along with the complete RMPs,
for other facilities and learn about their potential impacts, mitigation measures, and prevention
programs. The availability of this information would likely stimulate dialogue with local facilities
about differences in programs from one facility to another. Further, citizens may want to learn
about emergency response measures and capabilities at other communities associated with
alternative release scenarios to push for improvements in their own communities.
Ease of access has been shown to be vital for citizens and non-governmental
organizations. Environmental data sources that are easy to use see heavy usage, while those that
are harder to access do not. From July 1999 through January 2000, EPA's RMP*Info on the
Internet, which includes the RMP Executive Summaries, has logged over 155,000 page views of
RMPs. Note that RMP*Info does not currently provide OCA information; with OCA information
the usage would likely be higher. The "Scorecard" website from Environmental Defense
(formerly Environmental Defense Fund) likewise has found a high level of use. Scorecard
provides data on releases of Toxics Release Inventory (TRI) and other pollutants and data on
ambient air quality, interpreted for health risks and mapped for geographic sorting and display.
Scorecard has hosted over two million visitors since its launch in April 1998, and serves about
600,000 page views a month. The website also offers the ability for visitors to send a fax directly
to a company; about 5,000 faxes have been sent to about 3,000 distinct companies.3 TRI data,
easy to obtain on the Internet, have been used extensively by public interest and environmental
F-3
-------
groups, with the three most frequently reported uses being directly pressuring facilities for
change, educating citizens, and lobbying for policy changes.4
In contrast to the high usage of Internet websites, the public rarely accesses risk-related
data available only at state and local agencies that are not on the Internet. Information on
hazardous chemicals present at facilities and their quantities is provided to state and local agencies
under the Emergency Planning and Community Right-to-Know Act sections 311 and 312. In one
study, 41% of local emergency planning commissions (LEPCs) had no public inquiries during the
year of the study, and only 25% received over six inquiries.5
This low level of information transfer applies also to risk information in states where
facilities file reports, but where the information is not made available to the public except on
request. For example, New Jersey Bureau of Chemical Release Information and Prevention has
gathered information similar to RMPs for a number of years and makes paper copies available to
those who travel to its office or ask for them by mail; however, it has received fewer than 10
requests.6 Delaware has received similarly few requests.7 Nevada places its reports in public
libraries, but its experience indicates that librarians have to hunt for the files when updates are
needed. Few people have requested information from the state unless there has been an accident.8
Access is low, although these three states have had accident prevention program rules in place
since the late 1980s or early 1990s.
Why the disparity described above? The difference in usage is explained in part by what
researchers call the collective action problem. This problem occurs when a person chooses not to
take an action that would have net benefits to society because the costs to the individual of taking
the action outweigh the benefits to that person, or because he or she expects someone else to take
the action. Public information mitigates this problem by lowering these costs.9 Interest may be
substantial, but if access requires traveling across the county, requesting a specific document by
mail, or calling during the work day, many citizens will be unable or unwilling to gain access.
Making information publicly available helps overcome this problem.
Community and environmental organizations have expressed interest in comparing
facilities within a sector, within a region, from state to state, or within a company. For example,
in one case a refinery was up for sale and an environmental group sought to learn about the
environmental performance of companies that were potential purchasers. Non-governmental
organizations want to use data from OCAs to evaluate a facility's or company's hazard - in both
absolute terms and also, with national access to relevant data, in relative terms.10
The issue of public disclosure of hazard information was raised recently in Sacramento,
California. There, after the FBI arrested two individuals in an alleged plot to attempt to blow up
a large propane storage facility, the press reported on area residents' concerns. According to one
of multiple reports, "Subdivision residents plan to meet to discuss getting the propane tanks
moved or security increased, said Karen Banda, who bought a house there with her husband two
years ago. They asked the developers about the plant at that time, she said. 'If I remember
correctly, what they said was there was no concern,' Ms. Banda said. 'They should not have been
F-4
-------
allowed to build [housing] this close.'"11 Meanwhile, the facility's own RMP describes a far less
damaging worst-case scenario than reported in some accounts. The company's OCA yields a
worst-case distance to endpoint of 0.50 miles, based on one pound per square inch
overpressurization.12 Not only does this case show that OCA data are not necessary for would-be
criminals to target facilities with toxic or flammable chemicals (the facility made OCA data
available only later), but it reveals the value of public hazard information. Had this facility been
part of the RMP program (as it is now), and had the OCA information been easily accessible, then
residents - and potential residents - could have known what hazards existed in the area and acted
based on their level of concern.
Hazard information can raise concerns or it can lower them. Researchers at the Rutgers
University found that comprehensive outreach materials on accidental release scenarios not only
educates the public about risks and proper accident response, but it may actually serve to increase
trust and decrease worry.13 It is true that some facilities have voluntarily shared the OCA portions
of their RMPs and will continue to do so, and that OCA information does not address the
likelihood of a release. However, if members of the public could add widely available OCA
information to their understanding of accident likelihood (informed in part by RMP accident
histories), then the level of their concern likely would be in line with the apparent risks.
Equity is also a relevant consideration. A drawback of restricting access is that it will
likely have a disproportionate impact on poor and minority communities. RMP facilities may be
more often sited in poor, minority, or immigrant communities. If information is difficult to
acquire and interpret, then those with fewer advantages in income, language, or education are less
likely to be able to understand the hazards they face than those with more advantages. Traveling
to government offices, using raw data to construct OCAs, or hiring consultants to ferret out and
analyze highly technical information are all substantial hurdles. OCA information, since it is
interpreted and provides relevant information such as the distance that harm could potentially
extend, acts to overcome these hurdles. In addition, news media, citizen groups, and facilities can
bring OCA information to the people, counteracting to some degree the disparities in information
handling capacity. Granted, relying solely on the Internet for disclosure would also have
disproportionate impacts, at least in the short run. However, while a significant portion of the
public, including many in disadvantaged groups, do not have access to the Internet at home, many
gain access at work or at school, and the portion of the public with ready access is growing
rapidly.14 Other means, such as mail-out hard copies of OCA information, could be used to
address information needs for the time being.
What is the nature of access that would be used by citizens and citizens' groups for risk
reduction? Certainly, the public located near facilities handling hazardous substances will be most
interested in the OCA information and other RMP data for those local facilities. The statute
requires that the regulation provide for public access to a "limited number" of paper copies.
However, some localities have a large number of RMP facilities, and relying on paper copies and
read-only access means less use of OCA information.
Nature of Access That Would Be Useful
F-5
-------
For facilities in the local area, all OCA data elements, as well as the rest of an RMP,
would be valuable in better identifying and understanding local hazards. In addition, local
residents and local organizations desire information from OCAs for facilities with similar
processes or chemicals throughout the country, in order to allow comparisons. If someone were
to process OCA data into aggregations that happened to correspond to the needs of those living
near a facility and provide it to them, that facility could be compared with high, low, and average
values for a particular grouping, without facility identification. However, the lack of identification
information would prevent dialogue with facilities. Full OCA information (all OCA data elements
in a computerized database) would most efficiently serve the need for drawing comparisons.
Finally, national and state environmental and public interest organizations would be likely to use
all elements of OCA information for all facilities, for the same reason of allowing comparisons and
contacting facilities and companies of concern.
News Media
News media organizations, proven to be a large factor in the Toxics Release
Inventory reductions, have shown their interest in reporting on chemical
hazards. Were it more publicly available, the media would likely use OCA
information to gain easier access and obtain accurate, consistent information
set in context, on facilities both nearby and across the country. By reporting
on OCA information, the media will stimulate risk reduction by focusing the
attention of actors (communities, industry, and government) on the potential
off-site consequences of chemical releases and ways to prevent and mitigate
them.
Newspapers and other media have a strong interest and a long tradition in informing the
public about risks that they face, and what can be done about the risks. Importantly, coverage of
chemical hazards can raise awareness and motivation among many of the other parties discussed
in this chapter. This coverage stimulates action by stakeholders to engage in dialogue and reduce
risks.
Ample evidence supports the news media's interest and power regarding issues of
chemical release risk. Several papers have carried articles highlighting the-RMPs (without OCA
information) and hazards of local facilities. One example of press attention is a recent series by
the Washington Post. Following an article describing worst case scenarios of several local
facilities, the paper focused on concerns about the handling of chlorine at the Blue Plains sewage
treatment plant in Washington, DC. The plant had voluntarily put detailed OCA-related
information in its RMP Executive Summary, which is available widely. The paper described
scenario details (including a map showing the area potentially at risk), the health effects of the
chlorine and sulfur dioxide used at the plant, and apparent problems with plant safeguards. There
were subsequent articles over the next several days, and city officials, including the head of local
water authority and the mayor, took actions to improve safety monitoring, replace chlorine one
F-6
-------
and a half years earlier than previously planned, and beef up security. (Comparisons to OCAs of
similar facilities outside the region were not published.) 15
The media's effect in the Risk Management Program would likely resemble that in the TRI
program. There, a majority of respondents from citizen groups, state TRI agencies, and industry
agreed that media coverage of toxics issues has increased since the implementation of TRI.16
Newspapers have run articles listing specific companies or facilities as the top polluters in a town,
a State, or the country. Not only has coverage increased, but negative coverage has
corresponded with dramatic reductions in TRI releases; Appendix D provides a study of this
effect.
Newspapers and other media organizations have limited time and resources to research
stories; if easily accessible, OCA data likely would be used and chemical hazards reported more
completely. If, on the other hand, media organizations must obtain raw data from multiple
sources and then compile and analyze the resulting information about OCAs, the media would
likely report less. In addition, OCA data, as interpreted data that is already calculated and put
into a relevant context, would tend to be the most newsworthy part of RMPs. If it were available,
OCA data would likely attract the media's and the public's attention, first to the OCA data, and
then to other RMP elements.
News organizations have shown that they are willing and able to report on chemical
release risks, and this reporting clearly brings about risk reduction over time. However, while the
media are to some extent able to pursue their interests in hazard information locally, were
searchable OCA data available on a national basis, the media could more easily put hazards into
understandable terms for the public and draw comparisons among localities and facilities.
Nature of Access That Would Be Useful
In general, the news media are interested in accessing all OCA information on a national
basis. Certainly, local news organizations are most interested in information about local facilities.
However, access to all data elements and the ability to compare facilities with one another would
greatly enhance the ability of national and local media to put chemical hazards into context for the
public.
Emergency Planning and Response Organizations
F-7
-------
While local and state organizations responsible for emergency planning or response are
and will be entitled to OCA information, greater public disclosure carries advantages
for them and for risk reduction. These agencies often lack resources to fulfill their
planning tasks and often are concerned about penalties for improper dissemination of
OCA information. The greater the public accessibility to OCA data, the less the public
must rely on these agencies for its hazard information. Moreover, public access may
ease the planning and response burden for these entities because access will promote
better reporting by companies and more participation in planning and response by all
stakeholders.
OCA information and RMPs in general detail potential harm to the community (and to
responders). Planners and responders have responsibility to mitigate and reduce chemical release
risks to the community and to responders.
Local government officials will have access to OCA information as "covered persons" for
official use under CSISSFRRA. However, local organizations will often need some way to gauge
the level of risk and to evaluate the nature of their emergency preparedness and response. What
better way to gauge this effort than to use OCA information (worst-case scenario to some extent,
but alternative release scenario even more so) to identify locations that have similar problems in
order to learn about improvements in risk management or emergency preparedness and response?
Without broad OCA information availability, there will be no practical way to learn about the
passive and active mitigation measures used with the OCA assessments, and seeking out relevant
communities and facilities will be hit or miss.
Although state and local agencies are allowed access to OCA information for official use
as "covered persons" under the statute, there are two factors that restrict the further sharing of
information through these agencies. The first factor is that of resources. No stream of federal
funding is provided for LEPCs or SERCs, and they are often quite short of resources. In the
Mountain West, an estimated 75% of LEPCs have resource issues.17
Due partly to the resources problem, emergency planning and response agencies can
benefit, and information transfer can be enhanced, if information is available from a variety of
sources aside from themselves. For one, to the extent that hazard data could be obtained and
transferred without requiring local or state resources, the existing strain on resources is not
increased. In addition, broad dissemination allows public comparisons in a public forum, versus
dialogue only between facilities and covered persons. Broader dissemination would tend to bring
in outside viewpoints and issues and reduce the reliance on what are in fact often volunteer
organizations.
Regarding LEPCs' resources, some of the reasons public disclosure is beneficial have to
do with LEPCs' strengths and weaknesses. In general, LEPCs have not made a concerted effort
to bring hazardous materials issues to public attention, focusing instead on technical aspects.
F-8
-------
Furthermore, given the constraints under which LEPCs operate, it is unrealistic to expect LEPCs
to attempt to foster public debate of environmental issues or to focus on hazard reduction rather
than emergency response.18 A Government Accounting Office report noted that EPCRA section
312 chemical inventory data, which are provided to SERCs and LEPCs, are usually not even
computerized, and that use of the data by the broad public has been limited.19 Due to the resource
constraints, having information available from multiple reliable sources would reduce the reliance
on LEPC officials (and other covered persons), who otherwise would be responsible for
responding to more requests and questions with their limited resources. This, in turn, would lead
to both a reduced strain on agencies and greater public access to data of interest.
The second factor restricting information flow through state and local covered persons is
liability. CSISSFRRA includes a prohibition on unauthorized disclosure of OCA information by
Covered Persons. This prohibition, and the potential criminal penalties and fines of up to $1
million for violations judged to be willful, pose large disincentives for local and state officials to
pass on information involving OCAs. Many officials perceive that they are liable, and do not want
to be the arbiter of who gets access to what information. The statute continues to have a chilling
effect on officials obtaining and distributing information related to release risks.20
Government officials would likely make beneficial use of publicly available OCA
information, as they have other widely available information. For example, environmental agency
heads have publicly called upon the firms with the highest TRJ emissions to voluntarily reduce
their releases. In addition, for the state agencies that run TRI programs, the three most frequently
reported uses of TRI data are comparing the data to permits, source reduction efforts, and
comparing emissions patterns at similar facilities.21 Note that comprehensive comparisons are
possible only with access to nationwide data.
Responders and planners benefit indirectly from public access to the extent access
promotes compliance by industry and public involvement in planning. The additional scrutiny of
RMP compliance resulting from public access will lead to more complete compliance and higher
quality RMPs. This will, in turn, allow planners to use on these inputs and simplify planning.
Greater public involvement in planning will give planners more leverage in dealing with risks in
the community.
The drawbacks of restricting OCA information access and distribution include the
following. Local planners, fire departments, and others currently do not want to obtain OCA
information, in order to avoid liability if the same data were somehow released while in their
possession. Some local planners have indicated that because of the severe (Si million) potential
penalty, they would rather not take possession of the data, regardless of whether they are entitled
or have access.22 The perceived negative impacts have generated a chilling effect on the
desirability and use of OCA information and even other information associated with it.23
Nature of Access That Would Be Useful
F-9
-------
The types of access needed by emergency planning and response organizations are similar
to those that would be useful for citizens and citizens' groups. In general, for facilities within an
agency's area of responsibility, nothing less than all OCA information makes sense, although
alternative release scenario information is perhaps most useful to planners. This means that local
agencies would likely make use of local information and state agencies would use statewide
information. However, for facilities with similar processes, both local and state agencies can
benefit from disclosure of OCA information from such facilities nationwide.
Industry and Trade Associations
Companies with facilities subject to RMP requirements, which have the ability to
directly reduce the risk of chemical releases, have strong incentives to use OCA
information, including that from other companies, to efficiently reduce and manage
their risk. Producing high quality OCAs and accessing OCA data from other
companies would be most effectively achieved with a publicly available database.
This subsection describes the major incentives for risk reduction related to industries with
chemical release risks and their trade associations stemming from increased accessibility to OCA
information in a readily accessible form. In addition, this subsection discusses potential
consequences of restrictions of access to OCA information on industry's ability to prevent
catastrophic accidents.
The major categories of benefits or incentives for industry and trade associations
associated with the widespread dissemination of OCA information are:
• General Duty
• Potential for Greater Risk Reduction
• The Need for Continuous Improvement, and Demonstrating a Capability to
Manage Risk
Each of these categories are briefly described below.
General Duty
One of the most important prevention-related sections of the CAA is section 112(r)(l), the
General Duty Clause. This section provides that industry has a general duty - an obligation to
understand the hazards of their operations, design and operate a safe plant, take the necessary
steps to prevent accidents, and act to mitigate the consequences of those releases that do occur.
Consequently, satisfaction of this general duty means that a facility owner or operator must be
fully aware of all of the hazards present at its facility. In addition, the owner or operator must be
knowledgeable about, and make use of, hazard information, prevention practices, emergency
response practices, and all relevant industry codes and standards and regulations that apply to the
F-10
-------
industry. Full and complete RMPs and OCA information are essential tools that must be used to
the extent possible by industry as required by the general duty, and broad dissemination makes
using these tools more practical.
4
Potential for Greater Risk Reduction
Two key aspects are discussed: OCA information's critical role in process hazards
analysis; and how OCA information leads to risk management action.
At the heart of chemical accident prevention and process safety management is a formal
process hazards analysis, or PHA. A PHA involves the regular evaluation and identification of
hazards, assessment of risk and selection of risk control alternatives throughout the operating
lifetime of a facility. The majority of relevant information needed for a PHA is based on the
processes used at the facility. However, OCA information along with complete RMP details
assembled in one location for immediate use by hazard evaluation analysts across a number of
different industries and processes would broaden the process safety knowledge applied to PHAs.
A greater understanding of the wide variety of hazards, accident impacts, mitigation measures and
prevention programs serves to reduce the potential for catastrophic chemical accidents in specific
processes. The importance of various elements in a chemical accident prevention program is
often driven by the potential consequences of process upsets or failures. These consequences are
mirrored in the OCA information, especially in the alternative release scenario. Consequently,
suppressing OCA information from widespread publication would tend to diminish the overall
value and context of the RMP prevention information available to all industry and increase the
chance that information useful for a PHA was unavailable.
A thorough PHA along with hazard and consequence assessments is designed to reveal
vulnerabilities in a system that could lead to disaster. Once these vulnerabilities are revealed,
changes can be made to the chemicals, process technology or safeguards to address the
vulnerability. This leads to inherently safer processes.
The Risk Management Program and Risk Management Plan regulatory requirements are
flexible regarding facilities' equipment and operations. Although the rules provide a specific
framework, companies are obligated to assess for themselves the chemical and process hazards
and off-site consequences present at the site and to develop integrated accident prevention
measures and emergency response plans tailored specifically to the needs of the facility. The
facility bears the responsibility for evaluation, implementation, and documentation in the risk
management plan of the measures that will be used to protect the public and environment. With
this responsibility comes the burden of obtaining as much information as possible so that these
tasks are carried out as effectively and as accurately as possible. Widespread publication of the
plan for all to see drives a facility to make efforts that are more effective than under traditional
command and control regulatory approaches dictated by government. Only with publication of
complete RMPs along with OCA information, ideally in computerized form, can full use of RMP
information be made. Disclosure short of this would lead to reduced levels of effort.
F-ll
-------
In order to gauge the degree of consequences associated with their worst case scenarios
relative to their peers, companies must be able to compare their results with those of other
facilities. This could be done to some extent by using databases without identification information
or summaries showing data value ranges and averages (assuming someone would compile such
summaries). However, only with full OCA information in database form can a company not only
compare its performance with that of others, but use the identifying information to contact its
peers. Thus, publication of OCA information is shown to be important for hazards analysis.
The second important aspect of potential risk reduction is that OCA information leads to
risk management action.
EPA expects industry to react to public disclosure of OCA information in ways similar to
its reaction to publication of the Toxics Release Inventory. Industry reacted in very significant
ways to TRI. As related in Chapter 3, TRI emissions have decreased by 43 percent since 1988,
due to multiple factors. Case studies of eight companies and their response to the Emergency
Planning and Community Right-to-Know Act found that EPCRA "reporting and emergency
planning requirements have advanced the internal level of company awareness of their chemical-
related activities and releases, leading to the identification of areas for improvement and
subsequent action...[EPCRA] has resulted in greater transparency of company activities and [this]
has led to corporate action to reduce chemical risks."24 The two most frequently reported uses of
TRI data by industry are source reduction efforts and educating citizens.25 In one well-known
example, at the time of distribution of the initial TRI results, the Chief Executive of Monsanto
committed the company to a 90% reduction in TRI releases.
The evidence goes on. In one random survey of industry representatives, most
respondents reported that they were much more concerned with toxic releases since the passing of
EPCRA. The survey found a correlation between industries that said they had publicized
environmental problems or visible pollution and those industries that increased risk
communication with the public.26
Companies have increasingly joined voluntary programs intended to bring about change
without regulatory mandates but with public information. One example is EPA's 33/50 Program.
Launched in 1991, the program's goals included large reductions in releases of 17 TRI chemicals
from all reporting facilities. Measured relative to a 1988 baseline, the goals were for 33%
reduction by 1992 and 50% reduction by 1995. Although the exact extent to which the 33/50
Program caused changes in behavior is uncertain, 13,000 companies did participate and the
reduction goals were met ahead of schedule.27
Although there are differences in TRI and OCA information, EPA believes that the
disclosure of OCA information would lead to many risk management actions and a downward
trend in accidents and consequences similar to that of TRI releases.
The Need for Continuous Improvement and Demonstrating a Capability to Manage Risk
F-12
-------
A principle associated with good process safety management is the enhancement of
process safety knowledge and continuous improvement.28 Enhancement of process safety
knowledge involves internal and external research to "benefit from the latest advances in process
safety technology, and keep abreast of technological advances ..." (Ibid).
Even DuPont, long recognized as a world leader in safety, indicated that it has learned
from dialogue with citizens. "When a community resident and Citizen's Advisory Panel member
toured a DuPont facility after a recent release, [the member] suggested adding an ammonia
monitor at the top of a storage tank. The suggestion was judged to be a good one and DuPont
has since installed the device on the top of the tank."29 Clearly this example demonstrates the
value of local dialogue. However, the RMP and its related OCA information, data, accident
histories, and prevention measures provide a basis for this kind of dialogue nationwide.
Under Responsible Care®, member chemical companies are to: "continuously improve
health, safety, and environmental performance; listen to and respond to public concerns; and assist
each other to achieve optimum performance."30 (Responsible Care® is a chemical industry
initiative that is built around a set of six Codes of Management Practices. These codes include
Community Awareness and Emergency Response and Process Safety. Other elements of
Responsible Care®, which companies may choose to join, include a self-evaluation process that
determines how well companies are applying the Codes, mutual assistance, and performance
improvement measurements. See www.cmahq.com and www.socma.com/respcare.) Indeed,
changes at some facilities have been stimulated by a sharing of OCA information. In addition,
complete RMPs containing OCA information can be useful tools to identify facilities that are
safety leaders. Furthermore, small businesses with little technical resources and companies
evaluating opportunities for new business development can learn from current industry practices
to better control their future risks. Finally, chemical industry companies that are under
Responsible Care® "encourage and help other chemical manufacturers ... improve their own
performance through the Mutual Assistance Network, which provides direct interaction between
companies and the Partnership Program, which allows non-members to participate in Responsible
Care."31
Recently the chemical industry has worked to improve its public image. Part of this
openness is driven by Responsible Care® as noted above. In addition, companies have found that
dialogue is good for business. Certainly, making OCA information readily available for all to use
has image and marketing advantages for many businesses, and can be a good starting point for
dialogue with stakeholders. Many companies value their image as a responsible neighbor, and
realize that negative public relations can hamper efforts to expand operations or renew permits.
Note that Responsible Care® is specific to the chemical manufacturing industry. Although
Responsible Care® has been criticized as lacking in measurable goals, timelier, accountability and
credibility32, the many sectors that lack such an initiative may need even more help in addressing
risks. Access to complete RMPs (containing OCA information) would broaden the network of
assistance to the thousands of facilities that may not manufacture chemicals, but do handle them.
F-13
-------
Over time, companies would likely reduce the risk of releases by taking action due to
concerns brought to light by OCA information. These actions would reduce release hazards, the
likelihood of a release, or both. Note that many risk-reducing actions, particularly those that
reduce hazards, would reduce the risk of accidental releases and any risk that might be posed by
non-accidental releases. Examples of these actions include material substitution, reduction of
chemical inventories, passive mitigation measures, and emergency planning.
Some drawbacks to restricting OCA data access have been mentioned earlier in this
subsection. In addition, restrictions may allow important data to be withheld even from facilities'
neighbors, limiting the capability of local citizens to participate in risk management and
emergency response planning. Those interested in OCA information would likely find a "work-
around," and use other data and readily available tools (such as EPA's tool, RMP*Comp) to
calculate and publish their own versions of worst-case and alternative case scenarios instead of the
facilities' own. Such results are likely to not reflect conditions at the facility (e.g., process
quantity, passive mitigation, specific weather conditions), leading to misleading results and driving
attention to hazards that actually may be less significant and away from greater hazards.
Nature of Access That Would Be Useful
Companies and trade associations likely would make good use of several types of access
to OCA data, as follows. Individual companies could benefit from information on similar
processes from across the nation. Trade associations could make beneficial use of at least the
same.
Those with a Financial Interest in the Company
Those with a direct stake in a company's success at preventing releases have an
inherent interest in protecting that stake. Thus, to the extent OCA information is
easily available and covers similar facilities, insurers and others can use this
information to properly value their exposure and to pressure a company to improve
where needed.
Disclosure of OCA information can stimulate concerns about potential economic impacts
on those with a direct stake in a company with facilities handling hazardous chemicals. These
parties include insurers, investors, lenders, and those with other business arrangements with a
company. These interests are in addition to those of stakeholders in the community, described
below.
Insurers are interested in the ways companies manage risk. In general, they factor
indications of risk present into underwriting decisions and pricing. Regarding OCA information
and RMPs, conversations with insurance and loss protection corporation representatives reveal
F-14
-------
that these companies would make significant use of this information. As with accident histories in
RMPs, the details found in OCA information provide insurance and loss control companies with
greater understanding of the kinds of scenarios and potential losses that could occur, validating
the degree of loss control needed and validating expected insurance needs. Further, it would
enable the loss control industry to address customers' needs for assistance.33 Regarding process
industry and process safety, "what industry needs is to continue the development of performance-
based standards and regulations that follow the systemic approach and employ life-cycle models
that include the process itself, all safety/control equipment, and people (operators and
community). The approach must rely on risk metrics to support prudent business decisions ...",
according to one leader in the insurance industry.34 Disclosure of OCA information disclosure
would support for such community considerations and standardized risk metrics.
Another interested sector is comprised of investors and investment advisors and
researchers. These parties may disfavor purchase or ownership of company shares, and thereby
cause reduction in market value of shares. Interest is keenest in the field of "socially responsible
investing" (SRI). One of every eight dollars under investment management is now invested under
the rapidly growing umbrella of SRI, where an asset is evaluated along both traditional financial
measure and other measures.35 Those involved in SRI are very interested in environmental data,
and in comparing companies nationally.36 The level of interest among non-SRI investors and
portfolio managers lags behind.
Both mainstream and SRI fund managers often obtain data from research services. These
services tend to use data available from national databases, such as the Securities and Exchange
Commission's Electronic Data Gathering, Analysis, and Retrieval system (EDGAR) system. They
generally do not use data at the state and lower levels, due to the lack of consistency from one
data source to another, and due to the workload involved. Investment companies and research
services generally support enhanced access to information related to risk, but often only at the
federal level.37
Lenders comprise another interested group; they may be concerned that a potential
economic loss from an accident threatens a company's ability to repay loans, and also threatens
the value of assets (e.g., property) pledged as security for such loans. Other interested parties are
firms seeking to acquire a particular company, merge with it, engage in a joint venture with it, or
choose it as a primary supplier. They may be dissuaded from entering into such business
arrangements. This could be due to a perception that the company could not perform due to an
accident or its effects, or due to concern about incurring a share of accident liabilities under "joint
and several liability" or other liability-sharing doctrines.
Some organizations with a financial stake in the company perhaps would apply for access
to OCA information as qualified researchers. However, this would entail delay and restrictions,
and thus will be used less frequently than if OCA information were publicly available.
F-15
-------
It is clear that the parties described will have a stake not only in a company, but in its
hazard information. In general these parties are likely to seek OCA information from companies
with similar processes on a national basis.
Local Officials and Major Local Financial Stakeholders
Those with a stake in the community have an interest in the prevention of accidents,
and would likely use OCA information to learn about the hazards posed by local
facilities, and how those compare with those in other areas.
Local leaders who are not covered persons under the statute can benefit from public
disclosure of OCA information. Such leaders include government officials not in emergency
planning and response agencies (such as mayors) and persons with major financial stakes in the
community (such as landowners, real estate developers, and members of the chamber of
commerce). These interests are in addition to those of direct financial stakeholders in a particular
company, described above.
Disclosure can reveal risks to community safety, property, infrastructure, and natural
resources, and thereby prompt local leaders to use their powers and influence to prevent harms,
including the political and economic repercussions of an accident. In addition to concern over
risks stemming from an accident, there may be also be concern that information revealing large
and persistent hazards will itself cause economic loss (such as driving down property values).
Those affected by this would tend to take action to protect their interests.
The lack of knowledge about neighborhood hazards, as in the Sacramento case raised
above, applies not only to residents, but also to other public officials. For example, a school
board in Georgia hired a consultant to study the best locations for a new school building.
Available sites were near industry. OCA data were requested from industry, but withheld.
Consequently, data were derived from other sources and used in modeling; however a lack of
complete information caused the results to be misleading, generating significant land use planning
difficulties. Had the information been readily accessible, the costs and hassles that were caused
could have been avoided.38 As a practical matter, in any number of cases local entities will be
unable to obtain OCA data.
These local government and business leaders would likely make use of access to all local
OCA information. They may want to use data from outside their area to make comparisons, but
such data would be of less interest.
Workers and Labor Unions
F-16
-------
Those that work at facilities have a keen interest in knowing what hazards
surround them; OCA information would likely be used to raise awareness
and reduce off-site and off-site risks.
Although each company prepares its own OCA data, in general the data will not be
shared with all employees and contractors. Sharing of info across all industry would allow
workers and their unions to be more knowledgeable about hazards, safety controls, risk
management, etc. Unions have worked successfully to improve facilities' safety and
environmental performance, as noted in Appendix C. In addition, if a company were to provide
incorrect information in an OCA and it were made publicly available, those knowledgeable about
the facility could see and point out errors so that they could be corrected. In multiple cases, EPA
has had to work with companies to correct obvious errors in RMP submissions, but cannot
identify less obvious errors.
Since workers tend to live near facilities, they also would share the same concerns and
want the same data access as other local residents, noted above. Unions will tend to share the
same interests as research organizations, below.
1
Universities and Research Organizations
Universities and research organizations, by virtue of their educational and research
missions, have a natural interest in information that can lead to better understanding
the risks associated with chemical releases. The RMP and OCA information are
avenues to generate research around the world.
Multiple disciplines, including chemical engineering and risk communication, could be
advanced by easily available OCA information. Some institutions are particularly interested. For
example, the main mission of the Mary Kay O'Connor Process Safety Center, part of Texas A&M
University, is to improve safety in the chemical process industry.
Those in research organizations could certainly apply to be qualified researchers under
CSISSFRRA. However, certain aspects of this law will affect the ability of researchers to access
OCA information and publish findings based on it. First, the statute requires a system for access;
even the most efficient system will likely entail some delay. Second, the system "shall not allow
the researcher to disseminate, or make available on the Internet, the [OCA] information, or any
portion of the [OCA] information." (CSISSFRRA.) Although OCA information is narrowly
defined, this restriction can hamper the ability of a researcher to publish or support findings. The
likely effect of these factors will be less research into accident prevention and chemical security.
Thus, more immediate access and fewer restrictions on usage relative to the statutory minimums
are of interest to researchers, whether they are with a university, an industry trade association, or
a union.
F-17
-------
Universities and research organizations would likely make use of all OCA information data
elements across the country.
Professional Organizations
A number of professional organizations would likely use OCA information
to promote process safety.
Professional organizations have an interest in promoting process safety within their
professions. One example of an interested professional organization is the Center for Chemical
Process safety (CCPS), founded in 1985 following several chemical incidents. It is a division of
the American Institute of Chemical Engineers. CCPS commits itself to developing engineering
and management practices to prevent or mitigate the consequences of catastrophic events
involving the release of chemicals and hydrocarbons that could harm employees, neighbors and
the environment. Some areas of interest to CCPS sponsors include hazard and risk analysis,
engineering design, operations and maintenance, information dissemination and process safety
management.39
The CCPS has established a program to foster education about hazards analysis, process
safety, and emergency response at the undergraduate chemical engineering level. The Safety in
Chemical Engineering effort utilizes industry information in undergraduate chemical engineering
curricula so that graduating chemical engineers are better prepared to work safely in the chemical
processing industries. Course work includes problem sets and study of the concepts in hazard
analysis, consequence modeling, hazard evaluation, risk management, and process safety
management. This program could benefit from OCA information in conjunction with RMPs as
part of academic training and research.
In addition, the CCPS has produced numerous guidelines for process safety which rely on
industry information; future guidance could rely heavily on the hazard assessment and process
safety data contained in RMPs and OCA information. Further, CCPS conducts research into
technical problems and issues associated with hazards analysis and process safety.
Another example of interested professional organizations is the American Chemical
Society. It seeks, in part, to promote the public's understanding of chemistry and the chemical
sciences, and to foster communication and understanding among its members, the chemical
industry, the government and the community in order to enhance the quality of scientific research,
support economic progress, and insure public health & safety.40
The access needs for these and other professional organizations would match those of
universities and research organizations. That is, they would be interested in the most data
elements possible, and in the maximum geographic reach.
F-18
-------
Summary of Natures of Useful Access and Likely Impact on Risk Reduction
To characterize in even a very general way the different natures of access desired, one
must consider multiple variables. One variable is the geographic extent of information accessed,
which could range from the county to the country. Another variable describes the facilities of
interest. Interest in OCA information could range from all facilities to only one company, sector,
or process. The last variable covered here represents the data elements of interest. For example,
one party might make use of access to all data elements of an OCA, while another is content
without exact location or facility identification. For each segment of the public described in the
preceding section, Table 1 very briefly and generally notes the nature of interests in access to
OCA information.
Table F-l - Summary of the Natures of Interests in OCA Information
Type of organization/person
Geographic Extent
of Interest
Facilities of Interest
Data Elements
of Interest
Residents and Local Community,
Public Interest, and Environmental
Organizations
Local
All in local area
All (Individuals
less interested in
raw data)
National
Those with similar
chemicals or processes
All (Individuals
less interested in
raw data)
National and State Community,
Public Interest, and Environmental
Organizations
National
All
All
News Media
National
All
All
Emergency Planning and Response
Organizations:
• LEPCs and Fire
Departments
Local
All
All
National
Those with similar
chemicals or processes
All
SERCs
State
All in state
All
National
Those with similar
chemicals or processes
All
Industry and Trade Associations:
• Individual Companies
National
Those with similar
chemicals or processes
All
• Trade Groups
National
All
All
F-19
-------
Type of organization/person
Geographic Extent
of Interest
Facilities of Interest
Data Elements
of Interest
Those with Financial Stake in
Company
National
All
All
Local Officials and Major Local
Financial Stakeholders
Primarily local
All in local area
Distance to
endpoint and
receptors
Workers and Labor Unions
Local
Facility where working
All (Individuals
less interested in
raw data)
National
Those with similar
chemicals or processes
All (Individuals
less interested in
raw data)
Universities and Research
Organizations
National
All
All
Professional Organizations
National
All
All (Especially
mitigation used)
Which segments of the public would bring about the most risk reduction, given disclosure?
It is impossible to make predict for any particular case, but EPA believes that certain segments
will have a greater overall impact than others. The segments that would have the highest impact
on risk reduction are:
• Residents and community, public interest, and environmental organizations
• News media
• Emergency planning and response organizations
• Industry
Even though each company has access to its own OCA information now, industry and
trade associations would also have a significant impact if there were widespread disclosure. The
remaining segments of the public will likely have impacts, but they may not be as great as those
just mentioned. However, in any particular situation any type of person or organization could act
to significantly reduce risk.
F-20
-------
APPENDIX F REFERENCES
Fl. NICSNews. Vol. 8. (Fall 1999): p. 6.
F2. Harney, Kenneth R. "Web Sites Offer Helpful Information to Buyers and Owners."
Washington Post. (December 4, 1999): p. G12.
F3. EPA CEPPO; Correspondence with Dr. William S. Pease, Director, Internet Projects,
Environmental Defense Fund. Nov. 22, 1999.
F4. Lynn, F. and Kartez, J. "Environmental Democracy in Action: The Toxics Release
Inventory," Environmental Management, Vol. 18, No. 4, p. 511-521.
F5. Adams, William C., Stephen D. Burns, Philip G. Handwerk. Nationwide LEPC Survey:
Summary Report. The George Washington University (Oct. 1994).
F6. Baldini, Reggie. New Jersey Bureau of Chemical Release Information and Prevention,
personal communication with David Wiley, EPA, Dec. 8, 1999.
F7. Barrish, Robert. Delaware Department of Natural Resources and Environmental Control,
Division of Air & Waste Management, personal communication, with David Wiley, EPA, Dec.
20, 1999.
F8. Zusy, Mark. Nevada Department of Environmental Protection, Bureau of Waste
Management; personal communication with Craig Matthiessen, EPA, January 2000.
F9. Kleindorfer P., and E. Orts, "Informational Regulation of Environmental Risks. "Risk
Analysis. Vol. 18, No. 2: p.155-170. 1998.
F10. Epstein, Lois. Engineer, Environmental Defense Fund, personal communication with David
Wiley, EPA.
Fl 1. Lawrence, Steve. "News of plot has suburban residents rethinking industrial neighbor."
Associated Press, (December 6, 1999).
F12. Executive Summary, Risk Management Plan, Suburban Propane, Elk Grove, California.
(December 27, 1999) www.epa.gov/enviro.
F13. Kovacs, D., G. Gibson, C. Chess, and W. Hallman. "Outreach Materials About Risk
Management Plans: Guidance from Pilot Research." Cook College, Rutgers University Center for
Environmental Communication. (September 1998).
F14. "Falling Through the Net: Defining the Digital Divide." Department of Commerce, National
Telecommunications and Information Administration, July 8, 1999, revised November 1999,
www.ntia.doc.gov/ntiahome/digitaldivide.
F-21
-------
F15. "With Toxic Risk, Plans Vary," Washington Post, Oct. 10, 1999, p CI, C10-C11. "Plant
Warnings Go Unheeded," Washington Post, Nov. 5, 1999, p Al. "Blue Plains Details Safety
Plans," Washington Post, March 3, 2000, p B3.
F16. Lynn and Kartez, p. 517.
F17. Gablehouse, Timothy. Chair, Jefferson County Colorado LEPC, personal communication
with David Wiley, EPA, Dec. 2, 1999.
F18. Rich, R., David Conn, and W. Owens, "'Indirect Regulations' of Environmental Hazards
Through the Provision of Information to the Public: The Case of SARA, title III." Policy Studies
Journal, Vol. 21, No. 1, 1993: p 16-34.
F19. U.S. GAO, "Environmental Information: Agencywide Policies and Procedures Are Needed
for EPA's Information Dissemination," GAO/RCED-98-245. (Sept. 1998).
F20. Gablehouse, Timothy, Chair, Jefferson County Colorado LEPC, personal communication
with David Wiley, EPA, Dec. 2, 1999.
F21. Lynn and Kartez, p. 515.
F22. Gablehouse, Timothy. Chair, Jefferson County Colorado LEPC, personal communication
with David Wiley, EPA, Dec. 2, 1999.
F23. Howell, Kent. Georgia Department of Natural Resources, personal communication with
Craig Matthiessen, EPA, November, 1999.
F24. Baram M., P. Dillon, and B. Ruffle, Managing Chemical Risks: Corporate Response to
SARA Title III, Report to U.S. EPA, CEPPO, 1991.
F25. Lynn and Kartez, p. 515.
F26. Santos et al., "Industry Responds to SARA Title III: Pollution Prevention, Risk Reduction
and Risk Communication," May, 1994.
F27. Davies, Terry and Jan Mazurek. "Industry Incentives for Environmental Improvement:
Evaluation of U.S. Federal Initiatives," Resources for the Future, September 1996.
F28. Guidelines for Technical Management of Chemical Process Safety. Center for Chemical
Process Safety, American Institute of Chemical Engineers, New York, 1989.
F29. Dick Knowles in Chemical Manufacturers Association Summary of the Kanawha Valley
Safety Street Presentation, Charleston, West Virginia, June 1994.
F30. "Integrating Responsible Care, Community Advisory Panels, and Local Emergency Planning
Committees into Successful RMP Communications to the Public," Nick Macchiarolo, Great
F-22
-------
Lakes Chemical Corporation. Paper for CCPS conference 10/99.
F31. Ethyl Corporation, Health, Safety & Environment at www.ethyl.com.
F32. Baumann, J., P. Orum, and R. Puchalsky. Accidents Waiting to Happen: Hazardous
Chemicals in the U.S. Fifteen Years After Bhopal. (December 1999): p. 12-13.
F33. D'Angelo, C. Factory Mutual Corp., personal communication with Craig Matthiessen,
December 8, 1999.
F34. Stavrianidis, Paris. Factory Mutual Research Corp., "The 21st Century: Process Safety and
Factory Mutual," Mary Kay O'Connor Process Safety Center Roundtable Meeting, June 2-3,
1999. http://mkopsc.tamu.edu/symposiums/briefing%20papers/21st_century.htm.
F35. The Social Investment Forum 1999 Report on Responsible Investing Trends, Social
Investment Forum, November 1999, www.socialinvest.org.
F36. Gaskin, Russ. Managing Director, Social Investment Forum, personal communication with
David Wiley, EPA, Dec. 3, 1999.
F37. Bateman, Mark. Investor Responsibility Research Center, personal communication with
David Wiley, EPA.
F38. Howell, Kent. Georgia Department of Natural Resources, personal communication with
Craig Matthiessen, EPA, November, 1999.
F39. www.aiche.org/ccps
F40. www.acs.org
F-23
------- |