&EPA
United States
Environmental Protection
Agency
Office of Policy
Analysis (PM-220)
Washington, D.C. 20460
EPV230-06-89-061
December 1989
Acute Hazardous Events
Data Base (1989)
Final Report
•*«_ • .
*V-, • ••
Recyf.o Paper
-------
EPA-230-06-89-061
Acute Hazardous Events
Data Base (1989)
Final Report
Prepared for:
Office of Policy Analysis
Office of Policy, Planning and Evaluation
U. S. Environmental Protection Agency
Prepared by:
Industrial Economics, Incorporated
2067 Massachusetts Avenue
Cambridge, Massachusetts 02140
EPA Contract No. 68-W8-0038
December 1989
-------
ACUTE HAZARDOUS EVENTS DATA BASE
Table of Contents
FOREWORD ii
ACKNOWLEDGEMENTS iii
LIST OF EXHIBITS iv
EXECUTIVE SUMMARY I
1. PROJECT BACKGROUND AND DATA BASE CHARACTERISTICS 1-1
1.1 Proj ect Background 1-1
1.2 Methodology 1-3
1.2.1 Data Development 1-4
1.2.2 Data Editing 1-6
1.2.3 Development of Related Data Bases 1-6
1.3 AHE/DB Characteristics 1-8
1.4 Use of the AHE/DB 1-10
1.4.1 Source of Information on Specific Events 1-10
1.4.2 Perspective-Setting for Regulatory Actions....1-11
1.4.3 Perspective-Setting for Policy Decisions 1-11
1.4.4 Guidance for State or Local Actions 1-12
1.5 Cautions for Interpretation and Suggestions for
Additional Efforts 1-12
1.5.1 Completeness of the Data Base 1-13
1.5.2 Getting to the 'Truth — Cost/Quality
Trade-offs 1-15
1.5.3 The Problems of Causation, Consequences and
Priorities 1-17
1.5.4 Chemicals, Their Properties and Priorities....1-18
1.6 Report Organization 1-20
2 . CHARACTERISTICS OF THE MORE SEVERE EVENTS 2-1
2.1 Events Resulting in Deaths or Injuries 2-1
2.1.1 Extent of Deaths and Injuries 2-3
2.1.2 Circumstances of Release 2-4
2.1.3 Types of Facilities Involved 2-5
2.1.4 Chemicals-Involved 2-8
2.1.5 Relationship of Event Severity to Chemical
Properties 2-10
-------
2.1.6 Variation of Event Severity with Chemical
Properties 2-14
2.2 Evacuation Events 2-16
2.2.1 Extent of Damage 2-16
2.2.2 Circumstances of Release 2-17
2.2.3 Types of Facilities Involved 2-18
2.2.4 Chemicals Involved 2-20
2.3 Most Severe Events Involving Atmospheric Releases
of Acutely Toxic or Explosive Chemicals 2-22
2.3.1 General Characteristics 2-22
2.3.2 Significance of Event End Effects 2-24
2.3.3 Large Scale Domestic Releases of Extremely
Toxic Chemicals 2-24
2.3.4 Comparison with Major Industrial Disasters
of Similar Origin 2-26
3 . CHARACTERISTICS OF ALL EVENTS 3-1
3.1 Extent of Damage 3-1
3.2 Circumstances of Release 3-2
3.2.1 In-Plant Events 3-2
3.2.2 In-Transit Events 3-4
3.3 Types of Facilities Involved 3-5
3.4 Chemicals Involved 3-6
3.5 Geographic Patterns 3-7
APPENDICES
A. Data Coding and Format Sheets
B. Data Source Profiles
C. Chemicals in the ARE Data Base
D. Data Base Summary for All Events
E. Data Base Summary for Death or Injury Events
F. Data Base Summary for Evacuation Events
G. Data Base Summary for Events Reporting Spill/Vapor
Release or Fire/Explosion as Worst End Effect
-------
NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval does not
signify that the contents necessarily reflect the views and policy
of the Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
-------
FOREWORD
In 1985 EPA published criteria for selecting substances for
a listing of acutely hazardous substances. In addition, EPA
published the list resulting from the application of these
criteria. The data base that is the subject of this report was
originally developed as a part of the process of establishing this
listing criteria. Information from the data base was used in
drafting the published criteria. The results of this initial
effort, which involved the compilation of information on 3,121
records, are described in a December 1985 report, titled "Acute
Hazardous Events Data Base," EPA Report Number 560-5-85-029.
Since this initial effort, we have refined and supplemented
the data base. The data base now contains 6,190 records
representing 10,933 events. This represents a large share of the
most severe fixed facility events, and a large number of
transportation related events, which occurred in the United States
over the five-year period from 1982 to 1986. This data base has
been used both in the analysis of EPA policy options, and in
researching factors and issues related to the release of acutely
hazardous substances. Current plans include continued
supplementation of the event records contained in the AHE/DB,
verification of information on significant events, and the removal
of remaining duplicate event records.
ii
-------
ACKNOWLEDGEMENTS
This study was conducted under EPA Contract Number 68-W8-0038
by Industrial Economics, Incorporated. The principal members of
the project team were: Dr. James Cummings-Saxton, Ms. Amy J. Barad,
Mr. Marc Benoff, Ms. Janet A. Gochman, William C. Morrison, Dr.
Samuel J. Ratick, Mr. Robert E. Unsworth, Mr. Mark R. Veksler and
Mr. Anthony P. Wood.
The project team received valuable and generous assistance
from individuals in several organizations in acquiring and making
use of the data that went into this study. Among these
organizations are EPA's Office of Solid Waste and Emergency
Response and the Transportation Systems Center of the U.S.
Department of Transportation (DOT), which provided access to the
National Response Center's data base; U.S. DOT for access to its
Hazardous Material Information System; EPA's Region 7, which made
its computerized data base available; the Centers for Disease
Control of the Department of Health and Human Services; the
California Highway Patrol and Office of Emergency Services; the New
Jersey Department of Environmental Protection; the New York State
Attorney General's Office; the Ohio Environmental Protection
Agency; and the Texas Air Control Board.
Several EPA employees had active roles in the project. Mr.
Frederick W. Talcott has been responsible for the project since
its inception. Mr. Talcott has had a substantial role in
conceptualizing the data base, assembling and coding records,
reviewing interim products, and directing research using the data
base. Members of the Chemical Emergency Preparedness Program of
EPA's Office of Solid Waste and Emergency Response provided
suggestions and support throughout the project. EPA's Office of
Toxic Substances provided early funding, and individuals in that
office made valuable contributions. Members of the EPA Chemical
Accident Prevention Work Group have helped us shape the final
project.
111
-------
List of Exhibits
Exhibit Description
1.2-1 Data Sources for the AHE Data Base
1.2-2 Distribution of Events by Source and Year
1.2-3 AHE/DB's Three Interconnecting Data Bases
2.1-1 Occurrence of Injuries - All Events and D/I Events
2.1-2 Occurrence of Deaths - All Events and D/I Events
2.1-3 In-Plant Events by Location — All Events and D/I
Events
2.1-4 In-Plant Events by Cause - All Events and D/I Events
2.1-5 In-Plant Events by Event Context - All Events and
D/I Events
2.1-6 Frequency of Reported End Effects - All Events,
D/I Events and Evacuation Events
2.1-7 In-Plant vs. In-Transit Events - All Events and D/I
Events
2.1-8 In-Transit Events by Mode - All Events and D/I
Events
2.1-9 In-Transit Events by Cause — All Events and D/I
Events
2.1-10 Distribution of Events by Facility Type - All Events
and D/I Events
2.1-11 Reported Number of Deaths by Facility Type
2.1-12 Reported Number of Injuries by Facility Type
2.1-13 Frequency of Events Reporting Deaths by Facility
Type
2.1-14 Frequency of Events Reporting Injuries by Facility
Type
2.1-15 Frequency of Reported End Effects by Facility Type
- for D/I Events
iv
-------
List of Exhibits (continued)
Exhibit Description
2.1-16 Distribution of Events by Chemical for D/I Events:
10 Chemicals Representing the Largest Number of
Events
2.1-17 Substances Most Frequently Reported in D/I Events
2.1-18 Chemical Property Characterization Metrics
2.1-19 Frequency Distribution of AHE/DB Chemicals by
Chemical Property Ratings of Primary Substance
Released
2.1-20 Frequency Distribution of AHE/DB Events by Chemical
Property Ratings
2.1-21 Role of Fire/Explosion vs. Toxicity as Apparent
Cause of Deaths or Injuries
2.1-22 Event Severity by Toxicity Rating
2.1-23 Event Severity by Flammability Rating
2.1-24 Event Severity by Reactivity Rating
2.2-1 Occurrence of Injuries — All Events and Evacuation
Events
2.2-2 Occurrence of Deaths — All Events and Evacuation
Events
2.2-3 In-Plant vs. In-Transit Events — All Events and
Evacuation Events
2.2-4 Frequency of Reported End Effects - All Events and
Evacuation Events
2.2-5 In-Plant Events by Location - All Events and
Evacuation Events
2.2-6 In-Plant Events by Cause - All Events and Evacuation
Events
2.2-7 In-Plant Events by Event Context — All Events and
Evacuation Events
-------
List of Exhibits (continued)
Exhibit Description
2.2-8 In-Plant Events by Location — Total Amount Released
and Amount Released in Evacuation Events
2.2-9 In-Plant Events by Cause — Total Amount Released and
Amount Released in Evacuation Events
2.2-10 In-Transit Events by Mode — All Events and
Evacuation Events
2.2-11 In-Transit Events by Cause — All Events and
Evacuation Events
2.2-12 In-Transit Events by Mode - Total Amount Released
and Amount Released in Evacuation Events
2.2-13 Distribution of Events by Facility Type - All Events
and Evacuation Events
2.2-14 Frequency of Events Reporting Evacuations by
Facility Type
2.2-15 Percentage of Evacuees by Facility Type - Residents
Versus Workers
2.2-16 Substances Most Frequently Reported in Evacuation
Events
2.3-1 Air Releases of Toxic Chemicals Leading to Two or
More Deaths
2.3-2 Air Releases of Toxic Chemicals Leading to Forty or
More Injuries
2.3-3 Air Releases of Toxic Chemicals Leading to Two
Thousand or More Evacuees
2.3-4 Events with Highest Quantity-to-Toxicity Ratio
2.3-5 Ten Major Chemical-Related Industrial Disasters in
the Twentieth Century
3.2-1 Distribution by Release Size for All Events
3.2-2 In-Plant Events by Location — Total Amount Released
and Amount Released in D/I Events
VI
-------
List of Exhibits (continued)
•
Exhibits Description
3.2-3 In-Plant Events by Cause - Total Amount Released and
Amount Released in D/I Events
3.2-4 Distribution by Release Size for In-Plant Locations
— Events Involving Process Vessels
3.2-5 Distribution by Release Size for In-Plant Locations
— Events Involving Storage Vessels
3.2-6 Distribution by Release Size for In-Plant Locations
— Events Involving Valves or Pipes
3.2-7 Distribution by Release Size for In-Plant Locations
— Events Involving Disposal
3.2-8 Distribution by Release Size for In-Plant Locations
- Events Involving Heating or Cooling
3.2-9 Distribution by Release Size for In-Plant Locations
— Events Involving Vehicles Not in Transit
3.2-10 Distribution by Release Size for In-Plant Locations
— Events Involving Equipment Failure as Cause
3.2-11 Distribution by Release Size for In-Plant Locations
— Events Involving Operator Error as Cause
3.2-12 Distribution by Release Size for In-Plant Locations
— Events Involving Upset Conditions as Cause
3.2-13 Distribution by Release Size for In-Plant Locations
- Events Involving Fire or Explosion as Cause
3.2-14 Distribution by Release Size for In-Plant Locations
— Events Involving Arson or Vandalism as Cause
3.2-15 Distribution by Release Size for In-Plant Locations
- Events Involving Disposal as Cause
3.2-16 Distribution by Release Size for In-Plant Locations
— Events Involving High Pressure or Temperature as
Cause
3.2-17 Distribution by Release Size for In-Plant Locations
— Events Involving Power Fluctuations as Cause
vii
-------
List of Exhibits (continued)
Exhibit Description
3.2-18 Distribution by Release Size for In-Plant Locations
- Events Involving Sparking or Lightning as Cause
3.2-19 Distribution by Release Size for In-Plant Locations
— Events Involving Loading or Unloading as Event
Context
3.2-20 Distribution by Release Size for In-Plant Locations
- Events Involving Steady State Conditions as Event
Context
3.2-21 Distribution by Release Size for In-Plant Locations
- Events Involving Start-Up as Event Context
3.2-22 Distribution by Release Size for In-Plant Locations
- Events Involving Shut-Down as Event Context
3.2-23 Distribution by Release Size for In-Plant Locations
- Events Involving Maintenance as Event Context
3.2-24 Distribution by Release Size for In-Transit Events
Involving Trucks
3.2-25 Distribution by Release Size for In-Transit Events
Involving Rail
3.2-26 Distribution by Release Size for In-Transit Events
Involving Pipeline
3.2-27 Distribution by Release Size for In-Transit Events
Involving Barge
3.2-28 Distribution by Release Size for In-Transit Events
Involving Other Modes
3.2-29 In-Transit Events by Mode - Total Amount Released
and Amount Released in D/I Events
3.2-30 Distribution by Release Size for In-Transit Events
Involving Leaks as Cause
3.2-31 Distribution by Release Size for In-Transit Events
Involving Collisions as Cause
3.2-32 Distribution by Release Size for In-Transit Events
Involving Derailments as Cause
viii
-------
"List of Exhibits (continued)
Exhibit Description
3.2-33 Distribution by Release Size for In-Transit Events
Involving Over-Turned Trucks as Cause
3.2-34 Occurrence of Evacuations — All Events and D/I
Events
3.4-1 Distribution of Events by Chemical for All Events
— 10 Chemicals Representing the Largest Number of
Events
3.4-2 Substances Most Frequently Reported in All Events
3.5-1 Spatial Distribution of All Events
3.5-2 Spatial Distribution of Death Events
3.5-3 Spatial Distribution of Injury Events
3.5-4 Spatial Distribution of Evacuation Events
3.5-5 Spatial Distribution of AHE/DB Sources
.3.5-6 Spatial Distribution of Value of Shipments of
Manufactured Products (1985)
IX
-------
EXECUTIVE SUMMARY
PROJECT BACKGROUND
The Acute Hazardous Events Data Base (AHE/DB) has been
assembled as one component of EPA's ongoing effort to assess and
respond to the dangers posed to the U.S. public and workers by
sudden, accidental releases of toxic chemicals. The initial AHE/DB
contained over three thousand event records, which through data
sampling represented about seven thousand events. An interim final
report issued in December, 1985, titled "Acute Hazardous Events
Data Base" (EPA 560-5-85-029), characterized the 1985 version of
the AHE/DB.
Since 1985, the AHE/DB has been expanded by the Agency and
adapted for use in support of Title I and Title III applications
under the Superfund Amendments and Reauthorization Act (SARA) of
1986. The principal enhancements substantially expand the number
of records and augment the AHE/DB with supportive data bases
containing information on the chemicals and companies involved in
the reported events. These enhancements enable the AHE/DB to be
used more effectively to support the mandates of the Accidental
Release Information Program (ARIP) , the activities being undertaken
in the realm of Community Right-to-Know, and other policy analyses
conducted by EPA.
Because of EPA's primary interest in preventing events similar
to that which occurred at Bhopal, India in 1984, several criteria
were used to screen event data for inclusion in the AHE/DB.
Emphasis was placed on incidents that had acute hazard potential,
i.e., incidents resulting in deaths, injuries or evacuation were
given highest priority. In performing data collection for the
AHE/DB, events that resulted in releases of hazardous chemicals
were given priority, while those that involved the release of
petroleum-related products and non-hazardous substances were
largely excluded. Priority also was given to incidents that
involved air releases of chemicals. Events that involved the
release of large quantities of chemicals were given priority, and
events that involved the release of less than one pound of
materials were excluded.
AHE/DB CHARACTERISTICS
The AHE/DB is composed of four relational data bases. These
data bases are maintained on IBM AT-compatible personal computer
equipment. Two of these data bases (MAIN and SAMPLE) contain
information transcribed from existent chemical release event
reports. In addition, the study team developed two data bases
(CHEM and COMFAC) that provide information on the chemicals and
-------
facilities involved in events reported in the AHE/DB. The
information contained in these latter two data bases is derived
from a large number of additional sources that provide information
beyond that contained in the event reports. These data bases are
further described below.
The 1989 version of the AHE/DB contains 6,190 records, which
through sampling represent 10,933 separate events. The data come
from 38 sources. National sources include the National Response
Center (7,705 events), United Press International and Associated
Press (279 events), and the U.S. Department of Transportation's
(DOT's) Hazardous Materials Information System (168 events) . Other
sources include EPA's Region 7 (504 events), 26 daily newspapers
(327 events), and six offices of five state governments (1,820
events). Event records in the AHE/DB encompass the period from
before 1980 to 1987. However, the major data gathering effort
focused on the five-year period from 1982 to 1986. The AHE/DB
contains an average of 2,070 events per year for that period,
ranging from 1,501 events for 1982 to 2,619 events for 1984. Of
these events, 112 reported one or more fatality, and 968 reported
one or more persons injured. Altogether, 288 deaths and 10,803
injuries are reported in the AHE/DB. Of the 4,315 injuries for
which severity could be determined, 1,472 required hospitalization.
The Company/Facility Data Base (COMFAC) contains information
on facilities involved in AHE/DB events.1 This information
includes: facility name; address; a unique hierarchical identifier
containing the three-digit zip code for the facility and an
indication of any parent/subsidiary relationships; and a four-digit
Standard Industrial Classification (SIC) code for that facility.
All facilities in COMFAC were classified into one of five
categories: primary producers (facilities that produce chemicals) ;
secondary producers (facilities that use chemicals to manufacture
another product); distributors; end users; and waste-related
activities.
The Chemical Data Base (CHEM) contains information on the
chemicals involved in the events reported in the AHE/DB. This
information includes: chemical name; Chemical Abstracts Service
(CAS) number; four-digit SIC of the originating industry (i.e.,
the industry that produces the chemical); a commodity or specialty
chemical designation; and production characteristics. In addition,
CHEM contains chemical property information such as: toxicity
data, including "Immediately Dangerous to Life and Health" (IDLH)
values; physical characteristics, including vapor pressure, boiling
point, melting point, and air half-life estimates; and hazard
ranking information, including National Fire Protection Association
1For in-transit events, COMFAC contains information on the
company reportedly involved in the release (e.g., XYZ Trucking
Company).
-------
reactivity and flammability rankings. CHEM also identifies whether
or not the chemical is listed under RCRA, CERCLA or SARA (Title
III, section 302 and 313).
CHARACTERISTICS OF THE MORE SEVERE EVENTS
There exist a number of ways to analyze events. One way is
to consider the subset of the most severe events (i.e., those that
resulted in deaths or injuries), by reviewing the event
characteristics data presented in the AHE/DB. This section
demonstrates some of the attributes of the most severe events in
the AHE/DB.
Causes of Events at Fixed Facilities
Fire/explosion was the attributed cause more often in
death/injury (D/I) events at fixed facilities than for all events
at fixed facilities, accounting for one-eighth of the D/I events.
Equipment failure was the reported cause of more than one-third of
all events, and one-fifth of the D/I events. Operator error was
the reported cause in approximately one-eighth of all events and
D/I events.
Characteristics of Events by Event Cause
Fixed Facility Events
Attributed Percent of Death/. Percent of
Cause Injury Events All Events
Fire/Explosion 13X 4X
Equipment Failure 21X 36X
Operator Error 13X 14X
Facility Type
Each event location is assigned a facility type in the AHE/DB.
The first three facility types are primary producers (any facility
in SICs 28 or 29, i.e., Chemical and Allied Products, and Petroleum
Refining and Allied Products), secondary producers (including all
other manufacturing sectors, i.e., SICs 20 to 39, except 28 and
29), and end users (including residential and commercial users,
i.e., SICs 01 to 17, 48, 49 except 495, and 52 to 99). Although
events that occurred at primary producers rank second in terms of
the number of D/I events, they account for the largest share (39
percent) of events in which deaths occur. Events that occurred at
-------
distributors (i.e., events at wholesale, retail and transhipment
facilities), which account for the largest number of D/I events,
account for only eight percent of the total number of fatalities
reported in the data base. End users and secondary producers rank
second and third in terms of the number of fatalities, accounting
for 17 percent and nine percent of the deaths, respectively.
Events at facilities involved in waste disposal activities account
for two percent of the fatalities reported in the data base.
Twenty percent of deaths occur at facilities for which insufficient
information is available to determine facility type.
Although most events involving distributors are in-transit
events (69.8 percent), some of them occur at fixed facilities.
Similarly, while a large majority of events involving primary
producers (88.0 percent), secondary producers (94.0 percent), end
users (86.0 percent), and waste-related facilities (81.8 percent)
occur at fixed facilities, a small fraction are in-transit events.
Characteristics of Events- by Facility Type
Death/Injury Events
Facility Percent of Death/ Percent of Death Percent of Percent of Percent of
Type Injury Events Events Deaths Iniurv Events Injuries:
Primary Producers 26* 39X 43X 26X 31X
Secondary Producers 13X 15X 9X 13X 18X
Distributors 31X 12X 8X 32X 25X
End Users 13X 18X 17X 13X 9X
Waste: Disposal 1X 3X 2X IX IX
Unknown 15X 13X 20X; 15X 16X
* Total may not sum to 100 due to rounding.
Primary producers also rank first with regard to the number of
injuries—3,343 injuries occurred at such facilities, which
represents 31 percent of the total number of injuries reported in
the data base. Distributors rank second when injuries are
considered, accounting for 25 percent of all reported injuries,
while secondary producers account for 18 percent. End users and
waste disposers account for nine percent and less than one percent
of the injuries, respectively. Sixteen percent of the injuries
occurred at facilities for which insufficient information is
available to determine facility type.
-------
In-Transit Events
In addition to the primary cause of the event and the facility
type at which the event takes place, the' situation in which the
event takes place (i.e., fixed facility or in-transit) helps to
define the event. For example, in-transit incidents represent a
slightly larger percentage of D/I events than they do for all
events—37 percent versus 29 percent. Among in-transit events,
truck transport represents a significantly greater share (66
percent versus 49 percent) and rail transport a smaller share (28
percent versus 43 percent) for events resulting in deaths or
injuries, as compared to all in-transit events.
Characteristic* of Events by••;••'•
. •'•'•.••:'.:•. . • ln-Tnn»it Bocte . '.-:;.. ....".
Percent of Death/ Percent of
Mode Injury Events AH Events
Truck Transport 66X 49*
Rait Transport 28X 43X
Reportable Quantity
When the quantities of substances released in D/I events are
compared to their reportable quantities (RQs),2 35 percent of the
D/I events for which released quantities are reported involve
quantities less than the RQ. Note that the release of chemicals
in quantities less than their RQs would not usually be reported to
the National Response Center or other jurisdiction. Only 17
percent of the D/I events involved chemical release quantities
greater than 100 times the RQs of the chemicals involved.
2RQs determine those releases that must be reported under
section 304 of CERCLA. That is, all releases of substances in
quantities greater than their RQ would normally be reported to an
appropriate authority.
-------
Quantities Released in O/i Events
Qtlatiomnip to ffeporticl* Otantity
1-100 X (O C23.3IQ
rmn no c» an
1-10 X OOC33 ST)
Events Involving Evacuations
Evacuations were reported to have occurred in 829 of the
10,933 events represented in the AHE/DB. The 29 media sources (26
newspapers, 2 wire services and 1 trade journal) contributed 54
percent of the evacuation records. It is important to note that
the NRC data source, which is the primary source for 70 percent of
AHE/DB records, does not identify whether or not evacuations were
carried out. In the 405 events for which the number of evacuees
was identified, the total number of persons evacuated was 435,851.
Injuries were associated with 40 percent of the evacuation
events compared to only nine percent of all events. Deaths were
associated with five percent of the evacuation events compared to
one percent of all events. For some of the evacuation events it
was possible to distinguish between evacuations of workers and non-
workers. Of the 308,822 people evacuated in these events, 280,250
(91 percent) were residents and 28,572 (nine percent) were workers.
Characteristics of Evacuation Events
Percent of
Evacuation Events
Involving Injuries 40X
Involving Deaths 5X
Involving Deaths or Injuries 40X
-------
Events at facilities classified as distributors and primary
producers account for the majority of evacuation events (26 and 25
percent, respectively). End users and secondary producers account
for the next largest shares (14 and 13 percent, respectively) .
Compared to their frequency of involvement in all events,
distributors and end users have a higher frequency and primary
producers a lower frequency of involvement in evacuation events.
Secondary producers and waste disposal facilities account for a
similar proportion of all events (11 and one percent, respectively)
and the subset of events involving evacuations (13 and one percent,
respectively).
Characteristics or Evacuation Events
Fact Hty Type (Pixad f*ci Hty 6v«nt«)
OittriDutori (2V.0*3
Pr.nary a*oauc«r> C23-QK)
End LMTt C" 00
LARGE SCALE DOMESTIC RELEASE OF TOXIC CHEMICALS
Potential toxic effects vary both with chemical toxicity and
the quantities of the substances released. In the AHE/DB, 2,875
event records contain information on both quantities of material
released and acute toxicity. For those events for which we had the
necessary data, we divided quantity released by the IDLH value
attributed to that chemical to establish an empirical measure of
the potential toxic effects of each release. Using this measure,
25 AHE/DB event records involved releases whose quantity-to-
toxicity ratio was at least one-half of that associated with the
release at Bhopal, and 17 event records involved releases whose
ratio exceeded that at Bhopal.
-------
For the 25 events whose quantity-to-toxicity ratio was at
least one-half of that associated with the release at Bhopal, only
one event reported deaths (two fatalities), five events reported
injuries (ranging from five to 650 people injured) , and seven
reported evacuations (up to 20,000 people). The average numbers
of injuries and evacuations were somewhat greater for these 25
high-damage-potential events than for the AHE/DB as a whole, and
the average number of deaths was approximately the same. However,
given the vast quantity of materials released in the high-damage-
potential events, their severity remains surprisingly low.
The difference between the potential for human harm and the
actual consequences of these events appears to be due, in part, to
a combination of the following factors: (1) the chemical's
physical state at the time of release was liquid or solid, and much
of the material remained liquid or solid when released, or other
physical/chemical properties prevented widespread airborne
exposure; (2) few people were close to the point of release; (3)
weather conditions were favorable; (4) preparedness at the site of
release included procedures to mitigate the extent of release and
population exposure; or (5) evacuation or other emergency response
may have removed potentially exposed victims. In addition, the
data on these events are sometimes incomplete (e.g., the
concentration of some materials released, such as sulfuric acid,
may have been less than 100 percent). In most cases quantity data
represent "best estimates," which may overstate the actual quantity
released.
RELATIONSHIP OF EVENT SEVERITY TO CHEMICAL PROPERTIES
Acute chemical toxicity is represented in CHEM by an IDLH
value assigned to each chemical. Toxicity values were obtained for
approximately 29 percent of the 651 chemicals in the CHEM Data
Base. Releases of these chemicals account for 67 percent of the
events in the AHE/DB. The "most toxic" chemicals in CHEM (those
chemicals in the top one-third of toxicity for chemicals with known
IDLH in CHEM) account for more than 61 percent of the release
events in the AHE/DB. Thus, it appears that severity of outcome
has some relationship to toxicity.
Information on chemical toxicity and properties, combined with
event characteristics, was used to determine whether casualties
resulting from release events were attributable to the toxicity of
the chemicals released. Using this information, approximately 65
percent of the injuries and 13 percent of the deaths reported for
events in the AHE/DB were apparently caused by chemical toxicity.
In contrast approximately 35 percent of the injuries and 87 percent
of the deaths were apparently caused by fires or explosions.
8
-------
CAVEATS AND INTERPRETIVE GUIDANCE
The purpose of developing and maintaining the AHE/DB is to
provide a historical perspective on the types of accidental events
that release potentially hazardous substances in the United States,
the substances involved, the causative factors leading to their
release, and the end effects of these releases. The AHE/DB was not
intended to provide a precise estimate of the nation-wide frequency
of events, the quantities released, or their consequences. In
fact, information developed recently indicates that the AHE/DB may
significantly under-report less severe events (events that did not
result in deaths, or which resulted in very few injuries or
evacuations).
In attempting to interpret the reported data, some cautions
are in order. First, neither the study team nor EPA have performed
an independent review of the circumstances or characteristics of
the reported events, nor have they verified the information
contained in most of the event records in the data base. The data
base is best viewed as an accurate compilation and codification of
previously reported information on acute hazardous material
releases. A number of reviewers have cited instances where data
available through other sources (including EPA and company records)
conflicts with information contained in the AHE/DB. Where
possible, the AHE/DB has been adjusted in these instances.
However, for the vast majority of the events (greater than 96
percent) , the information contained in the data base remains an
unverified transcription of existent records.
Second, the term "injury" is used to refer to physical
manifestations ranging from temporary respiratory or eye irritation
treated on-site to critical injury resulting in prolonged
hospitalization. In an attempt to categorize the severity of
injuries, where possible we coded whether or not the injured
parties were hospitalized. While this parameter serves as a
surrogate for injury severity, it should be noted that not all data
sources used in developing the AHE/DB provided information on
hospitalization. Also, it is important to note that not all
hospitalizations are necessarily serious.
Third, direct evidence of causality rarely exists in the
contributing sources to the data base. Some of the deaths or
injuries reported to have occurred in association with a particular
event may not, in fact, have been caused by the release of a
chemical during the event. For example, about 13 percent of all
events with reported casualties list the cause as collision,
derailment or overturned truck. For these events it is possible
that some of the reported casualties may be the direct result of
the accident, not the release of a chemical substance during the
event. Whenever possible, reports of deaths and injuries resulting
-------
from the physical impact of a collision were removed from the data
base. However, it is not possible to identify how many of the
deaths and injury reports remaining in the data base fall into this
category.
Finally, in some cases the source of data being used to
document an event did not provide exact numbers of persons injured
or evacuated, or quantity of substances released. In these cases
standardized rules were used to quantify such terms as "a few,"
"about 50," and "a thousand barrels." While these rules will on
average yield reasonable estimates, in specific cases the estimated
values made may vary significantly from the true (albeit unknown)
value.
Thus, the nature and limitations of the AHE/DB must be fully
understood and carefully observed in deriving any conclusions from,
or in addressing any corporate or public policy concerns using, the
information contained in the data base. We suggest a careful
reading of the full AHE/DB report. Readers should apply this
historical data together with other information in setting
priorities or taking actions affecting subsets of chemicals,
facilities or circumstances for the purpose of accident prevention,
mitigation and emergency preparedness. Additional information
could include chemical, physical and toxicological characteristics
of released substances; national production and use
characteristics; quantities of substances commonly present during
production, distribution, use and disposal; as well as local
circumstances. Readers are encouraged to contact the EPA project
officer with any questions or concerns.
10
-------
1. PROJECT BACKGROUND AND DATA BASE CHARACTERISTICS
1.1 PROJECT BACKGROUND
The Acute Hazardous Events Data Base (AHE/DB) has been
compiled as part of the U.S. Environmental Protection Agency's
(EPA) review of the dangers posed to the public and industrial
workers by sudden, accidental releases of toxic chemicals. Concern
regarding such releases was heightened by the accidental release
of a pesticide intermediate that occurred on 3 December 1984 in
Bhopal, India, which killed over 3,000 and injured over 200,000
persons. Shortly thereafter, the EPA formed an Acute Hazards List
Workgroup to investigate the safety-related characteristics of U.S.
industry with regard to accidental chemical releases, and to
identify those chemicals that appear to represent unusually high
risks. In late May, 1985 the decision was made to assemble an
Acute Hazardous Events Data Base as part of the Acute Hazards
Analysis review process.
An important factor in the initial AHE/DB development effort
was the need for a rapid assemblage of data to support various
aspects of the analysis process. A team of three contractors was
formed, with Industrial Economics, Inc. (lEc) responsible for
overall project direction. The other two contractors were
Management Technology and Data Systems, Inc. (MTDS) and PEI
Associates, Inc. By mid-August, 1985, the AHE/DB contained 3,121
events, which through data sampling represented 6,928 release
incidents. An interim final report titled, "Acute Hazardous Events
Data Base," (EPA 560-5-85-029) was issued in December, 1985.
Copies of the data base were made available to the public at this
time.
Because of EPA's primary interest in preventing events similar
to that which occurred at Bhopal, several criteria were used to
screen event data for inclusion in the AHE/DB. Emphasis was placed
on incidents that had acute hazard potential, i.e., incidents
resulting in deaths, injuries or evacuations. In addition, events
1Paul Shrivastava, Bhopal: Anatomy of a Crisis (Cambridge,
Massachusetts: Ballinger Publishing Company, 1987), p. 65.
1-1
-------
that resulted in releases of hazardous chemicals were given
priority, as were incidents which involved air releases of
chemicals or large quantities of chemicals. Events that solely
involved the release of petroleum-related products and non-
hazardous substances were not included. Events that involved the
release of less than one pound of materials were excluded.2
Since 1985, the AHE/DB has been expanded to include 10,933
event records, and adapted for use in support of Title I and Title
III policy analyses under the Superfund Amendments and
Reauthorization Act (SARA) of 1986. The principal adjustments
included augmentation of the data base with data on substances and
companies involved in the reported events, and substantially
expanding the number of event records. These modifications enable
the AHE/DB to be used more effectively to support the mandates of
the Accidental Release Information Program (ARIP), in policy
analysis issues concerning Community Right-to-Know, and in other
programs being planned or implemented by EPA. Examples of further
modifications made to the AHE/DB include:
o Supplementing event records in the data base (where
information was available) with additional information
characterizing events, such as whether injuries resulting
from the chemical release required hospitalization;
distinguishing between workers and non-workers evacuated
as a result of a release event; and adding data fields
that further define the circumstances surrounding the
release events.
o Reviewing the events contained within the AHE/DB with
regard to the revised inclusion criteria. That is,
events solely involving the release of petroleum products
were excluded from the data base. As a result, 3,489
events involving polychlorinated biphenyls (PCB's), and
122 events involving the release of petroleum-related
products (e.g., kerosene, fuel oil) were transferred to
separate data bases. The purpose of this effort was to
focus the AHE/DB on events involving toxic chemical
releases with the potential for inducing acute health
effects on an exposed population.
In addition, commonly occurring incidences not reported in
the AHE/DB include: reports of oily substances or foul odors;
incidences not related to the release of a chemical (such as a
truck accident which does not result in a chemical release); and
instances in which only non-hazardous materials were involved.
1-2
-------
o Verifying and supplementing the information contained in
the AHE/DB for some of the most serious events, i.e.,
those events that led to deaths, a large number of
injuries or evacuations, or which involved the release
of a large quantity of a highly toxic substance.
Additional examples of policy analyses which have been performed
using the AHE/DB include:
o Conducting a number of analytical studies using the
expanded AHE/DB, including the development of descriptive
statistics on evacuation events, the appraisal of the
relationships between event characteristics and chemical
properties, and an assessment of the type of facilities
at which release incidents have occurred.
o Developing a structure to provide information to state
and local emergency response committees to help them meet
their responsibilities under Title III of SARA. This
includes the development of "Historical Performance at
Comparable Facilities" reports.
o Analysis of the scope and comprehensiveness of the
information contained in the AHE/DB, including assessing
the level of under-reporting and bias in the reporting
of events in the AHE/DB.
Section 1.4 of this report provides additional information on
analyses which have been conducted using the data base.
It is important to note that the data collection criteria used
in assembling the AHE/DB has purposely biased the data base toward
events with characteristics considered to hold the greatest
potential for sudden/ large scale harm to the populace due to the
release of hazardous chemicals. Thus, the data base should not be
viewed as being representative of the population of all acutely
hazardous release events, or of the types of outcomes resulting
from the release of hazardous materials.
1.2 METHODOLOGY
The primary purpose of the AHE/DB is to provide an historical
perspective on the characteristics of events involving the
accidental release of potentially hazardous substances in the
United States. These characteristics include the substances
involved, causative factors leading to a substance release, and the
end effects of a release (e.g., human health or property damage).
The AHE/DB is not intended to provide a basis for the precise
estimate of nation-wide frequencies of events, quantities released
1-3
-------
or the consequences of these events. Rather, the AHE/DB is
intended to provide an assessment of the general scope of these
types of events.
1.2.1 Data Development
Our intention in developing the AHE/DB was to convert
information describing the characteristics of a large number of
accidental releases into a form convenient for statistical and
policy analysis. In our initial data gathering effort we found
that there was no one data source which was sufficiently
comprehensive either in its breadth or its depth of coverage to
supply the overall perspective on accidental releases in the U.S.
This was our approach: (1) access large numbers of records from
several different types of data sources; (2) screen out event
records that were not relevant to the AHE/DB (such as events solely
involving petroleum products), as well as duplicate event records;
(3) take samples of certain categories of records;3 and (4)
transcribe the information, in a standardized format, into a
computerized data base.
More than twenty separate contributing sources were used in
developing the AHE/DB. (Exhibit 1.2-1) These sources differed in
the amount, quality and type of information provided. Emphasis was
placed on locating and using sources that held the potential to
yield a large number of usable records—such as DOT'S "National
Response Center Data Base", New Jersey's "Emergency Response/Air
Release Data Base", and New York's "Attorney General's Toxic
Accident Data Base"—or those data bases that provided information
on the more serious events, i.e., those that led to deaths,
injuries or evacuations.
Data coding involved organizing the information provided from
these different sources into the format specified in a single two-
sided coding sheet (an example of the coding sheet is provided in
Appendix A). The coding process organized the descriptive
information contained in the original sources into a formalized and
consistent set of categories. This format enables the generation
of summary statistics and cross-tabulations as presented in this
report. Since the AHE/DB represents a standardized transcription
3Events that did not involve deaths, injuries or evacuations,
but which were reported in the NRC data base, and which met all of
the AHE/DB inclusion criteria, were sampled at a frequency of one-
in-ten (i.e., one event out of every ten events in the NRC data
base was coded in the AHE/DB). Thus, through sampling, the 6,190
records in the data base statistically represent 10,933 events.
All of the analysis and exhibits in this report are based on this
weighted number of events.
1-4
-------
of information from a variety of sources, each of which provide
differing event information, comment sections are provided in the
data base to retain information that does not fit the coding
scheme. The comment fields are intended to prevent the loss of
important descriptive information about the release incidents.
Comment information is often a verbatim transcription of the text
included in the contributing source. In other cases, the coders
utilized the comment field to report a summary of the information
contained in more extensive event records. The 1985 coding format
was re-organized and expanded during the second data collection
effort, based on the experience gained in the initial coding
effort. This revision included the development of standardized
fields to contain information which would otherwise only be found
in the comment fields.
The following percentages quantify the general nature of the
availability of data from the AHE/DB:
o 87.5 percent of the data fields containing
information in a text format'(e.g., company name,
city or state) contain data.
o 71.5 percent of the data fields containing
categorical data (e.g., event context, quantity
units or end effects) and boolean fields (e.g.,
existence or absence of deaths, property damage or
impact on the environment) contain information.
o 58.7 percent of the numeric fields (e.g., number of
deaths, amount of property damage or total quantity
of substances released) contain information.
In addition, an average of one printed line of descriptive
information is found in the comment fields for each record, with
a range of zero to 14 lines of descriptive information presented.
Appendix A presents the instruction sheet utilized by data
coders in making consistent categorical choices. Coders were
trained professionals with backgrounds or experience in
engineering, chemistry, risk assessment or other relevant
disciplines. A data coding program was developed for this effort
that provided error-checking to prevent inappropriate information
from being entered into any one data field.
In order to maximize the accessibility of the AHE/DB to EPA
policymakers, as well as other interested parties, the AHE/DB is
maintained in dBASE III format in a PC environment.4
4dBASE III is a registered trademark of Ashton-Tate.
1-5
-------
1.2.2 Data Editing
Data editing is an on-going process, including: (1) assuring
that the coded records accurately reflect the relevant data source;
(2) combining information from two or more sources for the same
event; and (3) utilizing the results of in-depth research of events
in the AHE/DB to upgrade information provided for the more serious
events. However, neither the study team nor EPA has made an
independent review or verification of the information in most of
the records. Thus, we did not discover events or information about
events that were not previously recorded elsewhere. Neither did
we "second-guess" the information contained in the sources used.
Thus, the AHE/DB is best viewed as an accurate compilation and
codification of previously reported information on acute hazardous
material releases.
Since the 1988 Draft Final Report was completed, data editing
efforts have focused on (1) verifying the accuracy of coding for
events cited in this report; (2) identifying chemical name synonyms
and standardizing chemical names used in the data base; and (3)
identifying and removing duplicate event records for the same
event. As noted above, verification of events involves assuring
that the coded records in the AHE/DB accurately reflect the
information source(s) used. No attempt has been made to use
alternative sources to verify this information. The identification
and standardization of chemical name synonyms is intended to
improve the accuracy and usefulness of the data base, through
application of the most commonly used names for all substances
reported as released.
In order to fully assess the occurrence of duplicates in the
data base, the study team constructed an expert system designed to
identify duplicate records based on a series of decision rules.
These rules considered variables such as the date of the event,
event location, chemicals released and the facility name. This
system has enabled the team to remove over 114 duplicate records
thus far. While this is believed to represent the majority of
duplicate event records, undoubtedly some duplicates remain in the
data base, since many are difficult to identify due to missing or
incorrectly reported information.
1.2.3 Development of Related Data Bases
Two other data bases have been developed to augment the
AHE/DB. These data bases provide information on the facilities at
which events have occurred, as well as information on the
properties of chemicals involved in events reported in the AHE/DB.
Both of these relational data bases contain information beyond that
which is provided in the event records. Working from standard
1-6
-------
published sources, we created brief summary dossiers on the
facilities and the chemicals appearing in the event records, to the
extent they are specifically identified in any one event
record. Linking the AHE/DB to these other data bases facilitates
analysis of the relationships between event characteristics,
including end effects, with facility and chemical characteristics.
(Exhibit 1.2-3)
The first of these relational data bases is the Company/
Facility data base (COMFAC), which contains information on the
facilities reported to have been involved in AHE/DB events. This
information includes: the facility name, address, unique
hierarchical identifier containing the three digit zip code for the
facility and an indication of any parent/subsidiary relationships,
and the 4-digit Standard Industrial Classification (SIC) code for
the facility. Information on the number of employees is also
included for some of the facilities in COMFAC. COMFAC is linked
to the AHE/DB through the hierarchical facility identifier.
Facilities in COMFAC were organized into five categories:
primary producers (facilities that produce chemicals); secondary
producers (facilities that use chemicals to produce other
products); distributors (which includes events at wholesale, retail
and transshipment facilities); end users; and waste-related
activities. Note that while most events at primary producers,
secondary producers, waste-related facilities and end users are
fixed facility events, a small fraction are in-transit events.
Likewise, though most events involving distributors are in-transit
events, some involve fixed facilities.
The second relational data base is the Chemical Data Base
(CHEM), which contains information on chemicals involved in AHE/DB
events. This information includes chemical name, chemical abstract
service (CAS) number, the 4-digit SIC code of the industry that
produces the chemical, an indication of whether the substance is
a commodity or specialty chemical, as well as production
characteristics. In addition, CHEM contains information on
chemical properties such as: toxicity, including "Immediately
Dangerous to Life and Health" (IDLH) concentrations; physical
characteristics, including vapor pressure, boiling point, melting
point and air half-life estimates; and hazard ranking information,
including the National Fire Protection Association reactivity and
flammability rankings. This data base also identifies whether the
chemical is included on any of the following lists: SARA Title III
section 302 and 313, CERCLA Reportable Quantity, or the RCRA
Hazardous Waste list (Codes P or U).
1-7
-------
For both COMFAC and CHEM, care was taken to eliminate
redundancies in the identification of facilities and chemicals.
A unique facility name was agreed upon with all other spellings or
name presentations maintained for reference. Similarly, unique
chemical names and CAS numbers were established.
1.3 AHE/DB CHARACTERISTICS
o The AHE/DB currently includes 6,190 records, which,
through sampling, represent 10,933 events.
o The records included in the AHE/DB encompass the period
from before 1980 to the present. However, the major data
gathering effort focused on the five-year period from
1982 to 1986. The AHE/DB contains an average of 2,072
events per year for that period, ranging from 1,501
events for 1982 to 2,619 events for 1984. (Exhibit 1.2-
2)
o Of the events reported in the AHE/DB, 1,019 involved
deaths or injuries.
o Of the 10,933 events represented in the data base: 4,324
-^ involved primary producers; 1,522 secondary producers;
2,689 distributors; 1,239 end-users; 154 waste related
activities; and 1,005 occurred where the facility type
was not specified in the source reporting the event.
o Although most events involving distributors are in-
transit events (69.8 percent), some occur at fixed
facilities. Similarly, while a large majority of events
involving primary producers (88.0 percent), secondary
producers (94.0 percent), end users (86.0 percent), and
waste-related facilities (81.8 percent) occur at fixed
facilities, a small fraction are in-transit events.
o For events involving deaths or injuries: 267 involved
primary producers; 132 secondary producers; 317
distributors; 133 end-users; and 13 waste related
activities. Information was not available on facility
type for 157 D/I events.
o Of the 10,933 events represented in the AHE/DB, 8,444
involved commodity chemicals and 834 involved specialty
chemicals. For the remaining 1,655 events, either the
released substance was not specified, or if specified,
could not be classified as a specialty or commodity
chemical.
1-8
-------
The AHE/DB contains information on substances and
quantities released, event location, description of
causative factors and description of end effects. The
latter include: deaths, injuries, evacuations of workers
and non-workers, as well as property and environmental
damage estimates.
Records in the AHE/DB were gathered from 41 separate
sources, the most significant of which included:
The National Response Center data base (U.S.
Department of Transportation) which contains
information on hazardous substances released to air,
water and land throughout the U.S. This data base
includes information on in-plant and transportation
related releases. Through a stratified random
sampling scheme applied to the NRG data, the 2,962
records reported in the AHE/DB from this source
represent a total of 7,705 events.
News Wire Services (UPI and AP). The United Press
International and Associated Press provide
nationwide coverage of large events and events
having severe consequences. Events involving
evacuations, injuries and deaths commonly are
reported in this source.
Marsh and McLennan maintains a data base which
characterizes incidents included among the 100 worst
industrial accidents from 1956 to 1984 (rated by
property damage). One-tenth of these events are
transportation related. The remainder are split
evenly between chemicals processing and petroleum
refining.
EPA Region 7 maintains a data base of incidents
affecting all media in Iowa, Kansas, Missouri and
Nebraska.
A variety of regional newspapers were reviewed to
gather information, mainly on large events and
events having severe consequences. These sources
provide regional coverage with emphasis on events
in the relevant state.
A number of state sponsored reporting systems were
accessed for data. These included: the Texas Air
Control Board data base, which covers air releases
from manufacturing facilities in Texas; the Ohio
EPA data base, which includes information on
manufacturing and transportation related releases
1-9
-------
of hazardous substances in Ohio; the California
Highway Patrol and California Office of Emergency
Services data bases which emphasize transportation
related events and releases to land in that state;
and the New Jersey Department of Environmental
Protection and the New York State Attorney General' s
Data Bases, which focus on releases with potentially
significant human health or economic consequences
in these states.
The U.S. Department of Transportation Hazardous
Materials Information System, which tracks
transportation accidents involving hazardous
materials.
1.4 USE OF THE AHE/DB
The AHE/DB was initially designed for a limited purpose—to
set into historical perspective the list of Extremely Hazardous
Substances that EPA was developing in 1985. It performed that
function, in showing that this list was relevant to the kinds of
releases which occur, and the substances most often released in the
U.S. Of the more serious events (i.e., those with reported deaths
or injuries), almost one-half involved the release of one or more
substances on the original Extremely Hazardous Substances (EHS)
list, while another 30 percent of the events involved substances
with known fire or explosive hazards.
The AHE/DB has been applied to numerous additional analyses.
Some of these are summarized below.
1.4.1 Source of Information on Specific Events
Although the AHE/DB is not a census of events, but rather a
partial and informal sampling of event records, it does represent
an excellent source of data on a broad range of acutely hazardous
release events. For this reason, it has been used to identify
specific types of events for further study. Some examples include:
o EPA's Office of Toxic Substances used the AHE/DB to
examine ways in which the Toxic Substances Control Act
(TSCA) premanufacture review process might be modified
to focus more directly on the potential for accidental
releases of new chemicals.
o An industry trade association, in examining how
evacuations are performed in the U.S., used the AHE/DB
to identify candidate evacuation events for further
study. The researchers then conducted detailed case
1-10
-------
studies on about 50 of these events, using additional
documentary evidence, including interviews with
individuals involved in the events.
A firm specializing in training films that review
chemical release events and suggest appropriate responses
has used the data base to identify events involving
chemicals of concern.
1.4.2 Perspective-Setting for Reo^ilatory Actions
The data collection process used in developing the AHE/DB
created a sample of event records that do not fully represent the
population of acutely hazardous release events. However, while
recognizing the biases that resulted from our chosen data
collection process, it is reasonable to examine patterns within the
variables used to characterize events contained in the AHE/DB.
This report does so in its numerous figures and tables—by
examining substances, causation, event types and end effects. In
addition, EPA has used the information contained in the AHE/DB to
pursue aspects of potential regulatory or other program
implementation actions. For example:
o EPA's Office of Pesticide Programs examined the frequency
of chlorine releases from swimming pools and the
consequences of these releases. These releases were then
compared to other sources of chlorine releases in a re-
examination of the use classification of swimming pool
pesticidal substances.
o EPA's Office of Air and Radiation Programs examined the
environmental loadings and reported acute health effects
of a set of substances in considering the potential
regulation of various sources of air releases.
Specifically, the AHE/DB has been used to examine the
extent to which accidental releases contribute to human
health risks.
1.4.3 Perspective-Setting for Policy Decisions
In addition to aiding in perspective setting for regulatory
actions, the AHE/DB has been applied to policy analysis tasks.
Specifically:
o The AHE/DB was one element used in clarifying the extent
of accidental releases in the U.S. and their potential
for human health and environmental damage. In addition,
1-11
-------
it has been used by the internal EPA workgroup
implementing the new Community Right-to-Know program
under SARA Title III.
The data contained in the AHE/DB have been used by
offices in EPA attempting to examine priorities from a
comparative risk perspective, with the potential for
changing policy priorities within the Agency.
The AHE/DB has been used in setting the initial threshold
levels for the Accidental Release Information Program
(ARIP) as a means of focusing attention on the small
fraction of reported events warranting follow-up data
collection.
1.4.4 Guidance for State or Local Actions
EPA and lEc have done preliminary work in developing profiles
of the types and severity of releases occurring at various classes
of facilities. This work on "Historical Behavior at Comparable
Facilities" is preliminary in nature. As such, it must be
carefully understood prior to application. However, it is possible
that by aggregating records into logical groupings of events, the
data base may help local officials and others to understand better
the range of accident types that they may have to plan for, even
though the historical record present in the AHE/DB for a specific
area may be limited. We have supplied information to EPA regional,
state or local officials for this purpose. In addition, we have
discussed a more generalized profile format with individuals
involved in activities of the Massachusetts State Emergency
Response Commission (SERC), and several Local Emergency Planning
Committees (LEPCs) in Massachusetts.
1.5 CAUTIONS FOR INTERPRETATION AND
SUGGESTIONS FOR ADDITIONAL EFFORTS
Several aspects of the AHE/DB deserve particular notice. It
is necessary that the reader take account of the limitations of
the data base, in order to make the most appropriate use of the
data while avoiding misinterpretations. In each of the sections
that follow, we describe how the data base was compiled and how
this affects interpretation of the data. In addition, we outline
additional research tasks that are underway to assess better the
implications of these factors.
1-12
-------
1.5.1 Completeness of the Data Base
Initially, the project team hoped that one, or at most a few
sources of data would be sufficient to characterize the extent,
scope, and attributes of chemical accidents. However, it was
determined that no single set of sources was adequate to serve the
purposes of the study, either in identifying events or in
documenting the circumstances surrounding them. For example, we
found that events that were reported in the news media might not
be found in EPA or state data sources, while the majority of events
reported in EPA and state data sources attracted little media
attention. Among those events reported in more than one source,
we found that the type of technical information provided varied.
That is, one source might track the quantity of material(s)
released, but not the number of deaths or injuries associated with
release events. The result of this original scoping effort was the
development of a multi-source study design. This methodology is
similar to many epidemiological studies, in which private
physicians, public and private hospitals, clinics, school records
and other sources are tapped to estimate the number of cases of a
particular disease.
As we entered the second phase of the study (concluding with
this report), we developed strategies to increase the
representativeness of the AHE/DB for those types of accidental
releases which qualify for inclusion. The goal was to develop a
complete census of the more severe accidental releases, i.e., those
with any fatalities, with more than a few reported injuries or with
more than a few dozen persons evacuated. To accomplish this goal
we added more years of coverage from some of the original national,
state and regional sources. We also extended the scope of our
research effort by compiling information from a number of news
media sources (each of which provided varying regional coverage),
and reviewed documentation of acutely hazardous release events that
were developed by researchers in this field. As each additional
source of information was accessed, we noted that a higher
proportion of the events collected duplicated those already
identified in the AHE/DB, while a correspondingly lower proportion
of new events were added to the data base. This state of declining
marginal returns, expressed as the increasing cost per event
identified, suggested that we had achieved a reasonably complete
tally of the more severe events. In addition, we had increased the
total number of events in the data base to the point where
descriptive statistics developed from the data base held
significance for planned analyses.
However, it remains reasonable to question how complete the
AHE/DB is in recording information on the most serious events.
After completing the current version of the AHE/DB, we investigated
this question through intensive investigations in the states of
1-13
-------
Louisiana, Massachusetts, New York and South Carolina for the year
1986. This effort was carried out in cooperation with researchers
from the Centers for Disease Control in Atlanta, Georgia. It
involved a review of information on acutely hazardous events that
resulted in deaths, injuries or evacuations that were available
from local and regional news sources, various state agencies, EPA
regional offices, local police and fire departments, and local
officials with knowledge of severe chemical accidents. This
additional data collection effort identified a total of 185 new
death, injury or evacuation events for the year 1986. More
importantly, this additional data collection effort indicated that
the sources used to compile the AHE/DB may not have been reporting
a significant fraction of those chemical accidents which resulted
in human deaths, injuries or evacuations in that time period. The
information from this effort is currently under review, but our
initial conclusion is that progressively more intense searching of
event data sources will yield additional events with a moderate
numbers of casualties and evacuations.
This last point is worthy of further consideration. Within
the current AHE/DB, the ninety-eighth percentile of severity (i.e.,
events in the top two percent of severity) includes events with 79
or more persons injured, 15 or more persons killed or 8,000 or more
persons evacuated. The additional data collection effort
identified only four injury events and no death or evacuation
events that would have been in this top two percent range of
severity. The ninetieth percentile of severity (i.e., events in
the top 10 percent of severity) entails 24 or more injuries, five
or more deaths or 2,500 or more persons evacuated. The
supplemental data collection effort identified only five injury
events and no death or evacuation events of this magnitude. Half
of all death events, 61 percent of all injury events, and 59
percent of all evacuation events uncovered during the additional
data collection effort are at or below the median level of severity
for events in the AHE/DB (two injuries, one fatality and 250
persons evacuated). Thus, although more intensive data collection
is likely to uncover additional events which resulted in a serious
outcome, the majority of the most serious events were reported in
sources contributing to the AHE/DB.
The reader should therefore be aware that the AHE/DB does not
represent a complete census of events. Moreover, any attempts made
to achieve a complete census of acutely hazardous substance
releases will experience increasing costs per reported event. In
practical terms, a census may be infeasible for the time period we
have addressed. EPA is pursuing this topic further, first to
define the extent of under-reporting among the data sources used
in developing the AHE/DB; second to assess the relative severity
of the unreported events as compared to those in the data base;
and finally to assess the need for additions and adjustments to the
data base. Note that the advent of SARA Title III, with its
1-14
-------
increased reporting requirements, together with efforts by EPA and
others to increase compliance with existing reporting requirements,
and generally greater awareness of the existence and importance of
such requirements may have already combined to increase the
availability of more complete reports on events occurring today.
The extent to which this structural shift in the nature of
reporting has already occurred is a topic for additional
investigation.
1.5.2 Getting to the Truth — Cost/Quality Trade-offs
Because of the unique design of this study—acquiring and
coding large numbers of reports of accidental releases of
substances that were reported in sources which were easily-
accessible to the study team—we were able to compile a substantial
historical record (representing over 10,000 events) at a very
modest cost (between one-half hour to one hour of professional time
per event, including study design, data acquisition, entry, editing
and reporting). The study team is familiar with other efforts,
albeit with different goals from this study that have invested ten
to 40 times this effort, and of one investigation, involving 50
case studies, which involved about 100 times this effort per event.
Finally, several investigations of highly important or widely noted
accidents have had many months of professional effort devoted to
them—in some cases more than one-thousand times the intensity
allocated for each of the records in the AHE/DB.
Clearly it is possible to apply any given degree of effort to
the investigation of accident histories. With increased effort per
event, one is increasingly likely to develop a full
characterization of these events. But, with limited resources, the
scope of coverage diminishes. The reader should recognize where,
based on the study design, we chose to be on the continuum of
effort per event, and should be aware of the inevitable limits that
this choice places on interpreting the results. That is, this
study was designed to provide a broad coverage of accidental
chemical releases rather than the complete documentation of any
single event.
The reader should also recognize that even generous
expenditures per event still would leave sizable gaps in the data.
It would be fair to say that twenty times the effort produces less
than twenty times the information. In fact, the circumstances
surrounding many of the remaining events are so complex,
contentious or remote in time that, no matter how large the level
of effort dedicated to their study, substantial uncertainties will
likely always remain.
1-15
-------
To investigate, understand, and illustrate the difficulty
involved in determining the "true" circumstances surrounding an
event, given both time and expenditure constraints, we performed
an in-depth analysis of a relatively large and well-known incident
which occurred on 30 October 1987. The incident involved a release
of hydrofluoric acid and a mixture of hydrocarbons at a Marathon
Petroleum refinery in Texas which resulted in a number of injuries,
and the evacuation of individuals from the surrounding community.
Reports made within one hour of the initial release, both to the
NRC and to EPA's regional office, identified the main substance
released, gave no quantity, and reported no casualties. A report
made to a state authority within one day of the event listed an
amount released that was five times lower than the amount that the
facility later indicated to be the best estimate. On the day of
the event, a news media source reported a quantity released based
on an estimate of the capacity of the tank releasing the chemicals.
This amount was five times greater than the quantity later
indicated by the facility as having been released. A follow-up
report by government investigators about six months after the
accident, and a report by the facility made to the Accidental
Release Information Program one year after the event, represent two
historical references which contain "final" quantity-released
estimates.
The source used for this event by the AHE/DB was a media
report, as the official records were not available at the time this
event was coded into the data base. If the study team had access
to the NRC record and relied solely on the information contained
within it, the AHE/DB would have indicated that there were no
casualties or evacuations related to this event. There would also
have been no estimate of the quantity released. If we had used the
amount reported to a state authority one day later, the AHE/DB
would have under-estimated the quantity released by a factor of
five. The source we used to code this event apparently erred by
a similar factor, but in the opposite direction. After reviewing
a draft of this report, the firm owning the facility contacted the
study team, and disputed the volume estimate, which it indicated
to be too high by a factor of five. We have indicated, both in the
text and in the exhibits where appropriate, where potential errors
of a similar nature might have occurred. To reiterate, this
incident serves to illustrate the potential problems of relying on
any one single report as the final word on any particular event.
There are no practical solutions to the problems described in
this section, other than a very large expenditure of effort to
identify and correct such discrepancies. However, it is likely
that over-estimates and under-estimates both of release quantities
and other aspects of these events may tend to offset one another,
on average. It is also interesting to note, for this particular
event, a reduction in the quantity estimated as released changed
1-16
-------
its importance (ranked in terms of quantity divided by IDLH, as
presented on Exhibit 2.3-4) by only a very modest amount, given the
combined quantity and toxicity of the substances involved.
Proposed future research includes an in-depth search for
detailed information related to a subset of AHE/DB events. This
effort will most likely entail both a random sample of events from
the data base, as well as a census of all "significant" events
(based on some measure of number of deaths, injuries or evacua-
tions) . The results of this effort will provide an indication of
the overall level of the quality of the data both in the AHE/DB,
and for other reporting sources, such as the NRC.
1.5.3 The Problems of Causation. Consequences and Priorities
The circumstances surrounding most chemical accidents are at
least moderately complex, and could take several pages of text to
describe. Most records of incidental release events that are
readily available contain less than one-half page of text and data.
Attributing causation is often a difficult, if not impossible task
for many events. Releases may involve multiple pieces of
equipment, several individuals and several separate events. Thus,
although causation is ascribed based on the sources of data
utilized by the AHE/DB this factor should be interpreted carefully.
However, since information is available for a large number of
events within the data base, some conclusions can be drawn
concerning causation. One instructive example involves the role
of chemical toxicity versus that of fire and explosions and other
circumstances as causative factors of human casualties.
Information is presented later in the report that indicates that
fires or explosions appear to be responsible for the bulk of the
fatalities recorded in the AHE/DB, including those few events that
reported 10 or more fatalities. In contrast, the toxic properties
of the released substances seem to have caused the majority of the
injuries among the records.
One reviewer of a draft of this report asserted that, in
general, collisions were a more important cause of fatalities among
the transportation accidents than are chemical toxicological
properties. However, detailed review of the data seems to indicate
a more complex profile. More than 46 percent of the in-transit
events involving injuries or fatalities report "leaks" as the cause
of the event; 34 percent report either "collision", "derailment"
or "over-turned truck" as the primary cause of the event. The
remaining events reported the cause as either "other" or "unknown."
Thus, independent of an assessment of the count of fatalities, the
major fraction of serious in-transit events seem to be related to
situations other than collisions or wrecks. In addition, the vast
majority of the total number of injured persons were from those
1-17
-------
events which resulted in a half-dozen or more casualties each.
In fact, for some events the number of injuries run into the
hundreds. Thus, since it seems implausible that more than a few
persons would be injured by the impact resulting from a collision,
we conclude that properties of the chemicals involved in these
events were most likely responsible for the vast majority of the
injuries from these types of events. Thus, the potential for
exposure and risk (including death) to individuals near a
transportation released event clearly existed during many of the
events reported in the data base.
In a similar vein, one reviewer has asserted that there is no
evidence that links chemical releases at fixed facilities to
fatalities in the community, i.e., that all releases from
facilities that resulted in deaths involved only workers. Some of
the sources used in developing the AHE/DB include specific
information or other indications that the persons reported as
killed during an event were, indeed, workers. Most sources are not
as direct about this aspect, but we agree that little information
exists to indicate that any non-workers were killed by accidental
releases reported in the data base. However, it is important to
note that in those cases where it is possible to distinguish non-
workers from workers, non-worker evacuations outnumber worker
evacuations by almost ten to one. We have not made a similar
assessment for events involving injuries, but several events with
the largest counts of injuries involved individuals from the
surrounding communities. Thus, the potential for exposure and risk
(including death) to individuals outside the plant site appears to
have existed for many of the events reported in the data base.
The above discussion indicates that "causation" and "outcome"
are complex and often difficult to define. In addition, it is
misleading to focus solely on any one factor, to the exclusion of
the others. The examples given above suggest various additional
analyses that may further illuminate the full range of factors
related to acutely hazardous substance releases.
1.5.4 Chemicals. Their Properties and Priorities
With 927 unique substances reported as released in the data
base, not all of them could be equally important for policy
planners and researchers studying incidental releases of acutely
hazardous substances.3 In this report we have organized the
chemicals and the events in which they were released in a variety
of ways—by number of reported deaths, injuries and evacuations;
by number of events resulting in deaths, injuries or evacuations;
5Detailed information is available in CHEM for 70.2 percent,
or 651 of these substances.
1-18
-------
by their association with relatively large quantities of substances
released; and by the type of activity or plant at which the release
occurred. Displaying the data in these ways, we find that the
chemicals ranking highest in terms of these metrics were common to
many release events, while only a handful of events are associated
with chemicals which rank order 10 and below. Since only a few
chemicals represent a large percentage of the release events, it
is possible to ascribe undue importance to a particular substance
or circumstance, and conversely to implicitly downplay other
equally or more important chemicals or situations.
One aspect of this problem was raised by a group that reviewed
a draft of this document. Polyvinyl chloride (PVC) is a ubiquitous
plastic material, used in a multitude of solid shapes. When it is
involved in a fire, it burns and subsequently contributes chemicals
containing chlorine to the resulting smoke. Several events
involving PVC have been reported—four included reported casualties
and nine entailed evacuations. These facts raise the question as
to whether it makes sense to consider this widely-used material in
a manner similar to highly toxic gases or volatile liquids.
The answer to this question would depend on an assessment of
the physical characteristics of PVC or any other similar substance
in question. Questions which might be asked include: is it easily
flammable or does it release particularly hazardous substances when
reacting with oxidizers? Is there some circumstance surrounding
the burning of this material that results in severe or modest human
exposure potential? Are there common practices during the
production, distribution, storage or use of this material that
present some risk to human health or to the environment?
Related to this issue are broader questions regarding the
physical state of substances involved in accidental releases. It
seems reasonable to assume that gases have a greater potential to
reach humans and to cause potential harm than do liquids or solids.
Among all events, 4.6 percent of instances involved substances that
are solids at standard temperature and pressure, 77.4 percent
involved liquids and 27.7 percent involved gases (see Exhibit D-
1) .6 But, among the events with reported human casualties, solids
play a role in 12.2 percent of events, liquids in 73.6 percent and
gases in 31.8 percent (see Exhibit E-l). Thus, all three physical
states play significant roles in reported chemical accidents, with
solids having a relatively more important role in, events with
deaths or injuries than they do among the less serious events, and
liquids being much more likely to be involved in serious events
than gases. It is important to note that fires or explosions can
6 Note that these percentages sum to more than 100, since
events may involve the release of more than one substance in a
variety of physical states.
1-19
-------
effectively convert liquids and solids into gases, and can result
in changes in the chemical composition of the substances involved.
Additionally, the processing or use of materials at elevated
temperatures or pressures, or in the presence of other substances
may explain the ability of these substances to cause damage.
Thus, in general, it would appear incorrect to use the
"typical" physical state of a substance as the only measure of its
human exposure potential. Proposed research into this topic area
includes a detailed review of the accident histories of several
substances that are normally handled as solids or liquids, but
which have been involved in release events that resulted in a
significant number of deaths, injuries or evacuations.
In conclusion, we suggest a careful reading of the full AHE/DB
report. Readers should apply this historical data together with
other information in setting priorities or taking actions affecting
subsets of chemicals, facilities or circumstances for the purpose
of accident prevention, mitigation and emergency preparedness.
Additional information could include chemical, physical and
toxicological characteristics of released substances, national
production and use characteristics, quantities of substances
commonly present during production, distribution, use and disposal
as well as local circumstances.
1.6 REPORT ORGANIZATION
The remainder of this report is divided into two major
sections and seven appendices. The major sections are:
o Characteristics of the More Severe Events; and
o Characteristics of All Events.
The first section reviews the characteristics of events
involving deaths, injuries or evacuations. These characteristics
include the extent of property damage or human health impacts; the
circumstances of release (such as location, cause, process state
or transit mode); the types of facilities involved; the chemicals
released; and the relationship of event severity to chemical
properties.
*
The second section provides a similar overview of information
contained in the AHE/DB for all events, including those events not
involving deaths, injuries or evacuations. Frequently reported
chemicals, quantities released, media affected, environmental and
property damage, and evacuations are discussed.
1-20
-------
The seven appendices provide detailed background information
on the AHE/DB. The Data Coding and Format Sheets (Appendix A)
identify the information included -(although few records include all
of this information). Data Source Profiles (Appendix B) identify
the principal characteristics of the various sources of data used
in developing the AHE/DB. Appendix C provides a listing of all of
the chemicals included in the data base; the 4-digit SIC code of
the industrial sector that produces the chemical; the Chemical
Abstracts Service (CAS) identification number; and the number of
incidents, injuries and fatalities that occurred as a result of
events at which the chemical was released. Appendices D, E and F
provide summary outputs (referred to as "Encyclopedia of Data11) of
AHE/DB information for all events, for death or injury events, and
for evacuation events, respectively. Appendix G provides a similar
summary for those events that reported a spill/vapor release or
fire/explosion as the worst end-effect.
1-21
-------
2. CHARACTERISTICS OF THE MORE SEVERE EVENTS
2.1 EVENTS RESULTING IN DEATHS OR INJURIES
Each of the records coded into the AHE/DB were examined for
event severity, specifically for indications of human casualties.
Altogether, injuries were reported in 9.1 percent of the events
and fatalities in 1.0 percent. In 8.2 percent of the records in
the data base, insufficient information was available to determine
if injuries had occurred. For 5.9 percent of the records
insufficient information was available in the source used to
determine if any fatalities had occurred as a result of the release
event. (Exhibits 2.1-1 and 2.1-2) Although some of these records
may in fact represent events that involved injuries or deaths, it
is not possible to estimate the degree of under-reporting of deaths
or injuries without further research.
Some cautions are called for in interpreting the reported
casualty data:
o Reported Deaths or Injuries; Only a small percentage of
the information contained in the event reports used to
construct the AHE/DB have been independently verified,
either by EPA or its contractors. Thus, one should
interpret these data as "deaths or injuries that were
reported by one or more of the contributing sources."
o Causation; Direct evidence of causation is rarely
provided in the contributing sources to the AHE/DB. Some
of the deaths or injuries reported to have occurred in
association with a particular event may not, in fact,
have been caused by the event. For example, some
casualties may well have been caused by the triggering
accident that released a substance into the environment
(e.g., a truck driver whose injuries were caused by a
collision, and not the succeeding exposure to the
materials released). Whenever possible deaths that were
explicitly the result of factors other than chemical
toxicity, such as truck collision were removed from the
data base. However, it is not possible to identify the
2-1
-------
proportion of deaths and injuries that remain in the data
base that may have resulted from factors unrelated to the
release of a chemical substance.
o Range of Severity; Although some of the sources used in
developing the AHE/DB indicated in detail the type and
severity of the reported injuries, most sources did not.
To categorize the severity of injuries for those records
where information was not provided, we coded, where
possible, whether or not the injured parties were
hospitalized. This variable then served as a surrogate
for injury severity. While this categorization scheme
provides some information concerning injury severity, it
should be noted that not all data sources provided
information on hospitalizations, and that not all
hospitalizations are necessarily serious in nature. That
is, even in the case of hospitalizations, the severity
of an individual injury could range from temporary
respiratory or eye irritation to a more critical
condition leading to prolonged hospital treatment.
o Numbers of Injuries; Some of the sources used in
compiling the AHE/DB provide exact counts of injuries
whereas others (explicitly or implicitly) provide only
approximations. If a source quantified the number of
injuries this number was recorded in the database.
However, it was sometimes necessary to translate phrases
such as "a few" or "about 50" into specific counts of
injuries. Thus, the accuracy of the information
contained in individual records which make-up the data
base varies.
Although there is understandable interest in making nation-
wide estimates of the total number of deaths and injuries resulting
from accidental releases of acutely hazardous substances, this
study was not designed with that purpose in mind. Rather, the
objective was to understand better the nature, causes and effects
of accidental releases. Before drawing conclusions regarding
national totals, the reader should be aware of the caveats
presented above and consider the following factors:
o The 29 media sources used (26 daily papers, two wire-
services and one trade journal) contributed only six
percent of the total number of events. But, these events
represent 28 percent of the total number of death/injury
(D/I) events in the data base. Although these sources
together cover the period 1980 to 1986, no one source was
exhaustively searched for the entire period, and most
were systematically searched for four years or less.
(Exhibit 1.2-3 and Appendices D and E)
2-2
-------
o Although each of the 26 daily papers had a separate
regional focus, some also reported moderate to severe
accidents from other regions of the country. Thus, some
events appeared in two or more of the sources. Given
this, we believe that most of the events which received
national media attention during the period 1980 to 1986
are included in the data base.
o Some events involving human casualties appeared in only
one of the sources searched. This suggests that one
would find reports of other events of similar magnitude
both by increasing the "depth" of the data collection
(i.e., by including additional years of data for sources
already accessed), and by the "breadth" of the search for
events (i.e., by examining other media sources with
varying regional focuses).
o Some types of events may not be reported in the data base
due to the circumstances under which the events occurred,
such as an event involving casualties that was limited
to a work-place. ' Federal OSHA and state occupational
safety and health records have not been examined to
determine the extent of over-lap between those records
and the ones in the AHE/DB. Thus, it is not possible to
estimate the number of additional records that might be
identified by accessing these types of sources.
o Many of the sources used in constructing the AHE/DB
reported on events during the initial stages of the
incidents. In fact, in some cases the reporting party
was responsible for coordinating some aspect of the
emergency response effort. Such reports tend to be
preliminary in nature, are frequently incomplete in some
way, and in some cases, contain inaccurate information.
2.1.1 Extent of Deaths and Injuries
o One thousand and nineteen events (9.3 percent of all
events) involved deaths or injuries. The total number
of deaths reported was 288 and the total number of
reported injuries was 10,803. (Exhibit D-2)1
o Of the 6,434 injuries for which severity is known, 1,472
required hospitalization. (Exhibit D-2)
Exhibits using an alphabetic designation, such as "D-2",
refer to exhibits presented in the appendices.
2-3
-------
o For those events in which the number of injuries is
known, the average number of injuries was 11.2, and the
average number of injuries requiring hospitalization was
5.0. (Exhibit D-2)
o For those events in which deaths were reported to have
occurred, the average number of deaths was 2.6. (Exhibit
D-2)
2.1.2 Circumstances of Release
o Vapor releases were reported as an end effect in 63.3
percent of the D/I events, spills in 56.5 percent, fire
in 25.0 percent and explosion in 17.7 percent. (Exhibit
2.1-6)
In-Plant Events
For the population of all AHE/DB events, and for that
subset of events that involved deaths or injuries,
storage and process locations jointly accounted for one-
half of the reported events. Valves and pipes were the
second most frequently reported location for all events,
but ranked third among D/I events. (Exhibit 2.1-3)
Insufficient information was available to determine event
location for twice as large a percentage of the D/I
events as compared to all events. However, the
percentage of events at process vessels, storage
locations, vehicles not in-transit, heating/cooling and
other locations are nearly the same for all events and
D/I events. The percentage of events involving releases
from valves, piping and fittings is ten percent greater
for all events as compared to D/I events. (Exhibit
2.1-3)
Fire and explosion was reported as the attributed cause
of events almost four times as often for D/I events than
for all events, accounting for 5.2 percent of all events
and 19.1 percent of D/I events. Equipment failure was
the reported cause of slightly less than one-half of all
events (46.9 percent), but somewhat less than one-third
of the D/I events (30.9 percent) . Operator error was the
reported cause in approximately 18 and 19 percent of all
events and D/I events, respectively. (Exhibit 2.1-4)
The event context was reported as "steady state" in more
than one-half of all events, and approximately 40 percent
of D/I events. Maintenance operations and start-up were
2-4
-------
both reported twice as frequently as the event context
for D/I events than for all events. The percentage of
events for which insufficient information was available
to determine context was greater for D/I events
(approximately one-third) than for all events reported
in the data base (one-quarter). (Exhibit 2.1-5)
In-Transit Events
o In-transit incidents played a slightly larger role in
the subset of D/I events than they did for all
events~37.4 percent versus 29.0 percent, respectively.
(Exhibit 2.1-7)
o A larger share of events involving deaths and injuries,
as compared to all events, reported truck transport as
the in-transit mode (66.1 percent versus 48.6 percent).
A smaller share of D/I events, as compared to all events,
reported rail transport as the in-transit mode (27.6
percent versus 43.3 percent). (Exhibit 2.1-8)
o Collision was reported as the in-transit cause of events
twice as frequently for the subset of D/I events as
compared to all events. Leaks accounted for the largest
share of in-transit releases for both the subset of D/I
events and for all events (46 percent). (Exhibit 2.1-9)
2.1.3 Types of Facilities Involved
The industrial classification code of the facility (for fixed
facility events), or the company (for in-transit events) that was
reported to have been involved in the release event was identified
for 85 percent of the 1,019 events that reported deaths and/or
injuries. The study team assigned a 4-digit Standard Industrial
Classification (SIC) code to the facility or company reported to
have been involved in the release event, based on information
contained in the Dun and Bradstreet Reference Book. Facilities
were then further classified into five broad industrial groups:
Primary Producers. including Chemical and Allied Products, and
Petroleum Refining and Allied Products (SICs 28 and 29) ; Secondary
Producers. including all other manufacturing sectors (SICs 20 to
39, except 28 and 29) ; Distributors, those firms involved in
transporting, storing and distributing chemical products (SICs 40
to 47, 50 and 51); End Users, including residential and commercial
users (SICs 01 to 17, 48, 49 except 495, and 52 to 99); and Waste
Disposal (SIC 495).
2-5
-------
Note that while most events at primary producers, secondary
producers, waste-related facilities and end users are fixed
facility events, a small fraction are in-transit events. Likewise,
though most events involving distributors are in-transit events,
some involve fixed facilities.
o Of all D/I events, the largest share (31.1 percent) was
associated with distributors. Primary producers, one of
the areas of emphasis in most emergency response
planning, were associated with a slightly smaller share
(26.2 percent). Secondary producers and end users were
associated with equivalent shares of D/I events (13.0
percent and 13.1 percent, respectively). Waste disposal
facilities were associated with only 1.3 percent of the
D/I events. (Exhibit 2.1-10)
o Although primary producers ranked second among facility
types in terms of the number of D/I events reported in
the data base, events at these facilities accounted for
the largest share (43.4 percent) of the deaths that
occurred. While distributors were associated with the
largest number of events, these events accounted for only
7.6 percent of the fatalities. End users and secondary
producers were associated with 17.4 percent and 8.7
percent of the fatalities, respectively. Events
associated with waste disposal activities accounted for
2.4 percent of the fatalities. (Exhibit 2.1-11)
o When ordered by the total number of injuries reported in
association with events, events that occurred at primary
producers rank first among facility types—3,343 injuries
occurred at such facilities, representing 31.0 percent
of the total. Events that occurred at distributors
ranked second in terms of the number of injuries
reported, accounting for 25.3 .percent of all injuries,
while secondary producers accounted for 18.0 percent.
End users and waste disposers were the reported facility
type for 8.6 percent and 0.7 percent of the injuries,
respectively. Eighteen percent of all injuries occurred
at facilities for which insufficient information was
available to determine facility type. (Exhibit 2.1-12)
o The frequency at which fatalities occurred among all
events, by facility type, is shown in Exhibit 2.1-13.
Primary producers were the reported facility type for
43.6 percent of the 9,928 events for which facility type
could be determined, and of those events 45 (or 1.0
percent) resulted in fatalities. Events which occurred
at secondary production facilities demonstrate comparable
frequency of fatalities (1.1 percent of these events
reported fatalities) while only 0.5 percent of events at
2-6
-------
distribution facilities involved fatalities. While
events that occurred at end users and waste disposers
reported fatalities more frequently (1.6 percent and 1.9
percent, respectively), it is important to note that
these higher frequencies may be the result of biases in
the reporting of only the more severe events which
occurred at end users and waste disposers.
o Similar patterns are demonstrated when the frequency of
injury is considered, except that in this case
distributors were the most frequently reported facility
type (11.8 percent of all events reported at distributors
resulted in injuries). End users and waste disposal
ranked second and third in terms of frequency of reported
injury events (10.3 percent and 8.4 percent,
respectively). Events which occurred at primary
producers and secondary producers were less likely to
involve injuries (6.0 percent and 8.2 percent,
respectively). (Exhibit 2.1-14)
o Vapor releases were the most frequently reported end
effect associated with D/I events for all facility types,
except distributors. For events that occurred at
distribution facilities, spills were the most frequently
reported end effect (81.4 percent), and vapor releases
the second most frequently reported (57.4 percent).
Fires and explosions each accounted for 20 to 37 percent
of the D/I events for all facility types, except
distributors and waste disposers. Fires were reported
in only 13.3 percent of the events at distributors, and
explosions in only 7.9 percent. Fires were reported in
one-half of the events at waste disposal facilities.
(Exhibit 2.1-15)
It is important to note that the determination of facility
type is based on the identification of the two digit SIC of the
facility at which the release occurred. Thus, facility type is
only meant as a general indication of the industry in which the
facility is a part. For example, while all producers of chlorine
are in SICs 28, many consumers of chlorine are also in SICs 28 and
29. That is, not all events involving chlorine at facilities
reported as "primary producers" in the AHE/DB occurred at
facilities which produce chlorine. Thus, facility type
designations are only a general indicator of the primary activities
at a facility.
2-7
-------
2.1.4 Chemicals Involved
The 11 chemicals most frequently reported in association
with D/I events were reported as released in 43 percent
of such events (see Exhibit 2.1-16 for a graph of these
substances by number of D/I events).*
Chlorine was reported as released in the largest
number of D/I events (121, or 12 percent of D/I
events). Five deaths and 1,054 injuries were
reported in association with these events.
(Exhibits 2.1-16 and 2.1-17)
Ammonia (anhydrous), hydrochloric acid and sulfuric
acid were involved, respectively, in 66, 64 and 62
D/I events, together representing 19 percent of such
events, at which seven deaths and 1,524 injuries
occurred. (Exhibits 2.1-16 and 2.1-17)
The seven chemicals that were the next most
frequently reported in conjunction with D/I events
(ammonia, sodium hydroxide, nitric acid, toluene,
styrene, phosphoric acid and toluene diisocyanate)
accounted for 12 percent of D/I events. Forty-four
deaths and 1,291 injuries were reported in
association with these events. (Exhibits 2.1-16 and
2.1-17)
Chlorine was the most frequently reported chemical
associated with D/I events at primary producers,
secondary producers, end users and waste related.
Chlorine was reported as released for 22 percent of the
D/I events that occurred at primary producers; 15 percent
of D/I events at secondary producers; and four percent
of D/I events at distributors. (Exhibits E-7, E-9, E-ll
and E-13)
After chlorine, the next three most frequently reported
chemicals in association with D/I events at primary
producers were ammonia (anhydrous), hydrochloric acid and
sulfuric acid. Together these chemicals accounted for
54 events, or 20 percent of the total number of events
that occurred at primary production facilities. (Exhibit
E-7)
2Since any given event may involve more than one chemical,
there is overlap in the numbers of events, deaths and injuries
reported in association with more than one chemical. Therefore,
all percents or combined totals of events, deaths or injuries in
this section should be interpreted as upper bounds.
2-8
-------
o For events that occurred at secondary producers, the four
chemicals most frequently reported as released, in
addition to chlorine, were ammonia, ammonia (anhydrous),
hydrochloric acid and toluene diisocyanate. Together
these substances accounted for 28 D/I events, or 21
percent of the total number of events that occurred at
secondary production facilities. (Exhibit E-9)
o For D/I events that occurred at distribution facilities,
the three chemicals most frequently reported as released
were sulfuric acid, hydrochloric acid and ammonia
(anhydrous). Together they accounted for 74 events, or
23 percent of the total number of D/I events that
occurred at distribution facilities. (Exhibit E-ll)
o For the 21 chemicals reported as released in the largest
number of D/I events, the average quantities of substance
released per event are shown in Exhibit 2.1-17. Among
D/I events for the 11 most frequently reported chemicals,
sulfuric acid releases involved, on average, the largest
quantities (51,000 pounds), followed by nitric acid
(17,000 pounds), and sodium hydroxide and hydrochloric
acid (each at 11,000 pounds). These averages are largely
biased by a few extremely large events. When the largest
event of each substance is excluded, the largest average
quantities involve hydrochloric acid (9,000 pounds),
anhydrous ammonia, and sulfuric acid (6,000 pounds each) .
o When the quantities of substances released in D/I events
are compared to the reportable quantities (RQs) for these
substances, 36.1 percent of the D/I events for which the
quantity released was known, involved releases of
quantities of substances less than their RQs. It is
interesting to note that the release of quantities less
than the RQs would not typically be reported to the
National Response Center or other jurisdiction. (Exhibit
E-4)
o Only 14.3 percent of the D/I events involved the release
of a chemical substance in quantities greater than 100
times the RQ of the substance. (Exhibit E-4)
Again, it is important to understand the limitations of the
data base. Many substances are involved in a large number of
release incidences due to: (1) the high volume of the substances
used in commerce or industry, or (2) their association with more
hazardous substances. For example, sodium hydroxide occurs
frequently in release events reported in the data base. However,
sodium hydroxide is not comparable in terms of toxicity with many
2-9
-------
of the more hazardous substances found in the data base. The large
number of events involving sodium hydroxide undoubtedly reflect
the ubiquity of this substance in industry and commerce, and
possibly reflect its association (and subsequent co-release) with
more hazardous substances.
2.1.5 Relationship of Event Severity To Chemical Properties
Information on five physical and toxicological properties—
acute toxicity, flammability, reactivity, volatility and processing
mode—is contained in CHEM, for those chemicals associated with
events reported in the AHE/DB. These data were collected from
various sources, which are identified in the subsections below.
In addition, a determination of the physical state of the released
chemicals was made based on information contained in the original
event reports for most of the records in the AHE/DB. Exhibit
2.1-18 summarizes the type of information used and the rating
scales employed for each property. Note that the following
analyses are based on the chemical properties of the primary
substance reported for each event.
Acute Toxicity
The acute toxicity of many of the chemicals reported as
released in the AHE/DB are characterized in CHEM using values
compatible with "levels of concern" (LOG) as defined in EPA's CEPP
Technical Guidance Document.3 Where available, National Institute
for Occupational Safety and Health (NIOSH) measures designated by
the term "Immediately Dangerous to Life and Health (IDLH)" were
used as a measure of acute toxicity. Additional IDLH values were
drawn from the CEPP guidance document.
For purposes of this analysis IDLH values are grouped into
five LOG categories, termed LRATEs, which are shown in Exhibit
2.1-18. The higher LOG indices correspond to hazard categories
previously established by organizations attempting to segregate
chemicals with regard to the hazards attending their usage. LRATE
4 corresponds to EPA's "Extremely Hazardous Substances" criteria
for categorizing chemicals for reporting under SARA Title III
section 302. LRATE 3 refers to "Hazardous Chemicals" subject to
SARA Title III section 311 and 312 material safety data sheet
^.S. EPA, Chemical Emergency Preparedness Program, Site
Specific Technical Guidance for Hazards Analysis; Emergency
Planning for Extremely Hazardous Substances. Draft, August 1987.
2-10
-------
(MSDS) and chemical inventory requirements. An LRATE value of 2
is equivalent to the SARA Title III section 313 "toxic" reporting
category. LRATEs 1 and 0 designate chemicals in the least toxic
categories.
o LRATEs were determined for approximately 29 percent of
the 651 chemicals in CHEM. (Exhibit 2.1-19)
o Releases of chemicals for which LRATEs were determined
accounted for 66 percent of the events in the AHE/DB.
(Exhibit 2.1-20)
o Forty-one percent of the events in the AHE/DB involved
chemicals in the highest toxicity group (LRATE 4) ,
although only 10 percent of the chemicals reported as
released in the AHE/DB fall into this toxicity category.
A slightly larger proportion of chemicals involved in
D/I events are in the highest toxicity group (19
percent). (Exhibits 2.1-19 and 2.1-20)
o Relatively few of the events involving deaths or
injuries, or of all events reported in the AHE/DB,
involved chemicals whose toxicity places them in the
LRATE 3 category. This is generally consistent with the
lower frequency of LRATE 3 chemicals in the CHEM data
base. (Exhibits 2.1-19 and 2.1-20)
Flammability
Information on the flammability of chemicals was obtained
primarily from the National Fire Protection Association (NFPA),
which rates substances for flammability (referred to in this report
as "FRATE") on a scale of 0 to 4, with FRATE 4 being the most
hazardous level.4 In cases where NFPA ratings were not available,
flash points and boiling points were used to assign similar ratings
based on a scheme developed by the National Academy of Sciences
(NAS). When neither NFPA ratings nor physical property data were
available, FRATEs were assigned according to NFPA criteria, for
cases where the necessary information was available.
National Fire Protection Association (NFPA), Fire Protection
Guide on Hazardous Materials. 1984.
'National Academy of Sciences (NAS), System for Evaluation of
Bulk Water Transportation of Industrial Chemicals: A Report to the
Department of Transportation. U.S. Coast Guard. 1974.
2-11
-------
o Information on flammability was obtained for
approximately 65 percent of the chemicals in CHEM.
Releases of these chemicals represent 87 percent of all
events in the AHE/DB. (Exhibits 2.1-19 and 2.1-20)
o Forty percent of all events involved non-flammable
substances (FRATE 0) , and approximately 30 percent
involved chemicals in the two highest flammability rating
groups (FRATEs 3 and 4) . The distribution of D/I events
by flammability rating is similar to the distribution for
all events, with the exceptions that a larger percentage
of these events report FRATE as "unknown," and a smaller
percentage report the highest flammability category.
(Exhibit 2.1-20)
Reactivity
Reactivity was characterized using the NFPA definition, "the
susceptibility of materials to release energy either by themselves
or in combination with water." Observed substance responses to
fire exposure, shock and pressure are used to determine such
susceptibility. As with flammability, NPFA assigns degrees of
reactivity (RRATE) on a scale of 0 to 4, with 4 being the most
reactive.
o Reactivity ratings (RRATEs) were determined for only 40
percent of the chemicals in CHEM. Very few of the
chemicals were assigned to the highest reactivity
category; only 4.0 percent were assigned RRATEs of 3 or
4. A similar pattern holds for D/I events, although in
this case RRATEs could be identified for 56 percent of
the released substances. (Exhibit 2.1-19)
o A comparable pattern is reflected when considering the
number of events, with only 3.2 percent of all events and
2.0 percent of D/I events involving chemicals with
RRATEs 3 and 4. (Exhibit 2.1-20)
Volatility
Vapor pressure, which indicates the tendency of molecules to
escape from liquids or solids, is used as a measure of volatility
in CHEM. Ratings of high, moderate and low volatility (VRATES)
were assigned to the chemicals based on criteria established by
2-12
-------
previous investigators (identified in Exhibit 2.1-18).6 These
categories are compatible with the chemical selection criteria used
for the SARA Title III section 305(b) Review of Emergency Systems.
o Vapor pressures have been identified for only 33 percent
of the chemicals in CHEM. These chemicals were reported
as released in one-half of all AHE/DB events. The rated
chemicals are evenly divided into the low, moderate and
high volatility groups. (Exhibits 2.1-19 and 2.1-20)
o A slightly greater percentage of the events (for which
chemical vapor pressure is known) involved chemicals in
the low volatility group. A similar pattern was
demonstrated for events involving deaths and injuries.
(Exhibit 2.1-20).
Physical State
A chemicals's physical state at the time of an event can be
an important factor determining the severity of an event. The
AHE/DB contains information on physical state, specifically:
o Information regarding the physical state of the primary
substance released was available for 99 percent of the
events reported in the AHE/DB. (Exhibit D-l) Note that
many events involve the release of multiple substances,
sometimes with different physical forms. Thus,
percentages sum to greater than 100.
o Releases of liquids predominate for both D/I events and
all events, representing the chemical state in 72 percent
of D/I events and 74 percent of all events. (Exhibits
D-l and E-l)
o Gases are the second most commonly reported physical
state, reported in 32 percent of D/I events and 27
percent of all events. (Exhibits D-l and E-l)
o Solids were reported as released much less frequently,
—only 12 percent of D/I events and four percent of all
events reported the release of solids. (Exhibits D-l and
E-l)
Versar, Inc., Physical-Chemical Properties and Categorization
of RCRA Wastes According to Volatility, prepared for the U.S. EPA,
February, 1985; and U.S. EPA Office of Toxic Substances, Section
305 (b) Review of Emergency Systems — Chemical Selection Criteria.
Draft, January, 1987.
2-13
-------
Processing Mode
As a proxy for processing mode, chemicals reported as released
in the AHE/DB are designated as either commodity or specialty
chemicals. This proxy is based on the assumption that commodity
chemicals, which are produced in large quantities, are likely to
be produced in continuous processing modes, whereas specialty
chemicals, which are produced in small quantities to meet more
narrowly defined specifications, are more likely to be batch
processed.
o Approximately 41 percent of the chemicals in CHEM are
commodity chemicals, 35 percent are specialties and 24
percent could not be designated. (Exhibit 2.1-19)
o Seventy-seven percent of all events involve commodity
chemicals, while only seven percent involve specialty
chemicals. Although the shares of D/I events involving
commodity and specialty chemicals are similar to the
shares for all events, the distribution of chemicals
involved in D/I events is more heavily weighted to
commodities. (Exhibits 2.1-19 and 2.1- 20)
2.1.6 Variation of Event Severity With Chemical Properties
Acute hazardous events can be separated into two groups
depending upon whether or not a fire or an explosion occurred
during the event. In those cases where a fire or explosion was
reported as an end effect, the severity of the event, in terms of
number of injuries or deaths, appears to be predominantly
determined by the characteristics of the fire or explosion. For
those events where a fire or an explosion was not reported as an
end effect, we assume that an important factor contributing to the
severity of the event was the toxicity of the chemicals involved.
Since multiple end effects are often reported for a single event,
individual records were examined in order to consider all
information when classifying the cause of the deaths or injuries.
Variation With Toxicity
o Approximately 65 percent of the injuries and 13 percent
of the deaths reported in association with events in the
AHE/DB were apparently caused by toxicity. (Exhibit
2.1-21)
2-14
-------
o For the subset of events that resulted in deaths and
injuries which were thought to be caused by chemical
toxicity, more than two-thirds of the events, for which
a chemical toxicity rating could be determined, involved
highly toxic chemicals (LRATE 4). (Exhibit 2.1-20)
o Events involving chemicals with high LRATEs had a much
wider range of severity (in terms of numbers of deaths
and injuries) than those with a low LRATE. (Exhibit
2.1-22) A chi-square analysis of the grouped data
suggests that this phenomenon for D/I events is due to
the underlying distribution of the two variables in the
data set. That is, statistical analysis indicates that
there may be a relationship between severity (as measured
by the number of deaths and injuries) and level of
chemical toxicity.
o Chi-square analysis on the toxicity ratings of chemicals
involved in all events in the AHE/DB, for which the
primary end effect was a vapor release or spill, suggests
that a relationship exists between event severity and
chemical toxicity. The analysis indicates that severity
is related to toxicity in determining whether or not
deaths or injuries will occur, but that the degree of
severity, i.e., number of deaths or injuries, does not
depend on toxicity. That is, if an accidental release
of a highly toxic chemical occurs, the event is more
likely to cause deaths and/or injuries. However, we
cannot predict the number of deaths or injuries.
Variation of Events Severity
With Flammability and Reactivity
o Approximately 87 percent of the deaths and 35 percent of
the injuries reported in association with events in the
AHE/DB were apparently caused by fire or explosion.
(Exhibit 2.1-21)
o A chi-square analysis indicates that there is no apparent
relationship between the severity of events, as measured
by the numbers of injuries plus deaths, and the
flammability or reactivity ratings of the chemicals
involved. (Exhibits 2.1-23 and 2.1-24)
2-15
-------
2.2 EVACUATION EVENTS
The expanded coding format for the present version of the
AHE/DB allowed each record to incorporate information on whether
or not an evacuation was performed, how many people were evacuated,
and of those, how many were residents and how many were workers.
Evacuations occurred in association with 829 (or 7.6 percent) of
the 10,933 events in the AHE/DB. Of the remainder, 5,067 (or 46.4
percent) of the events did not have associated evacuations, while
for 5,037 events (or 46.1 percent) it was not reported whether an
evacuation took place. The 29 media sources (26 newspapers, 2 wire
services and one trade journal) contributed 54 percent of the
evacuation records.
Similar caveats to those discussed for death and injury events
apply to this analysis, particularly regarding the number of
persons evacuated and the distribution among residents and workers.
2.2.1 Extent of Damage
o Over 800 events (or 7.6 percent of all events) were
reported to involve evacuations. In the 405 events for
which the number of evacuees was identified, the total
number of people evacuated was 435,851. (Exhibits D-l
and D-2)
o Injuries were associated with 39.8 percent of the
evacuation events, compared to only 9.1 percent of all
events. (Exhibit 2.2-1) Deaths were associated with 4.8
percent of the evacuation events compared to 1.0 percent
of all events. (Exhibit 2.2-2)
o For some of the evacuation events it was possible to
distinguish between the evacuation of workers and
non-workers. Of the 335,323 persons evacuated in these
events, 308,183 (91.9 percent) were residents and 27,140
(8.1 percent) were workers. (Exhibit D-2)
o For events where the number of persons evacuated was
reported, the average number of evacuees per event was
1,076. For those events in which a distinction was made
between residents and workers, the average number of
residents evacuated was 973 and the average number of
workers evacuated was 99. (Exhibit D-2)
2-16
-------
2.2.2 Circumstances of Release
o In-transit events played a slightly larger role in
evacuation events as compared to all events—34.4 percent
of all evacuation events versus 29.0 percent of all
events. (Exhibit 2.2-3)
o Spills were reported as an end effect in 49.9 percent of
evacuation events, vapor release in 76.1 percent, fire
in 32.3 percent and explosion in 15.1. Fires and
explosions were reported nearly five tines more
frequently in association with evacuation events than for
all in-plant events. Vapor releases also were more
frequently reported, and spills less frequently, as end
effects in evacuation events as compared to all events.
(Exhibit 2.2-4).
In-PIant Events
o Storage locations were reported to be involved in 35.1
percent of in-plant evacuation events, compared to 28.1
percent for all events. Vehicles not-in-transit
accounted for 5.7 percent of evacuation events and 2.8
percent of all events. Process locations accounted for
a smaller share of evacuation events (19.5 percent) when
compared to all events (22.4 percent), as did valves and
pipes (14.3 percent for evacuation events versus 25.4
percent for all events). (Exhibit 2.2-5)
o Fires and explosions were reported as the attributed
cause four times more frequently in evacuation events
(21.5 percent) when compared to all in-plant events (5.2
percent). Sparking and lightning, and arson and
vandalism were also more frequently reported as causes
in evacuation events. Equipment failure represented the
event cause in 46.9 percent of all events, but only 33.3
percent of evacuation events. (Exhibit 2.2-6)
o Steady state condition was the in-plant event context in
over one-half of all events (51.0 percent) and evacuation
events (55.9 percent) . Start-up conditions were reported
as the in-plant event context three times more frequently
for evacuation events (4.4 percent) than for all events
(1.8 percent), while loading and unloading was reported
as the in-plant event context one-half as frequently (9.6
percent) for evacuation events as for all events (16.2
percent). (Exhibit 2.2-7)
2-17
-------
Events involving releases at disposal locations accounted
for 83.0 percent of the total quantity of substances
released during in-plant evacuation events, as compared
to 5.9 percent of the substances released during all
in-plant events. The large share of material released
at disposal locations during in-plant evacuation events
was attributable to a single event which released a very
large quantity of substance (16,500,000 pounds of
hazardous waste). (Exhibit 2.2-8)
Fire and/or explosion, which was the attributed cause for
5.7 percent of the total quantity of material released
in all in-plant events, accounted for 84.0 percent of the
material released during in-plant events which involved
evacuations. (Exhibit 2.2-9) This aberration results
from the same, large single release of 16,500,000 pounds
of hazardous waste.
In-Transit Events
o Truck modes were responsible for 50.2 percent of
evacuation events while 45.3 percent involved rail. This
represents approximately the same percentages as their
shares in all events. (Exhibit 2.2-10)
o Derailment was the attributed cause in 24.9 percent of
the evacuation events, which is approximately four times
more frequent than derailment's share of all events (6.5
percent). Collisions were the attributed cause in 21.4
percent of the evacuation events, which is almost twice
the share of all events accounted for by collision (11.6
percent). (Exhibit 2.2-11)
o Rail-associated releases accounted for two-thirds of the
quantity of substances released in in-transit evacuation
events. However, rail related events represented only
22.3 percent of the quantity of substances released in
all in-transit events. (Exhibit 2.2-12).
2.2.3 Types of Facilities Involved
o Of the 829 events for which evacuations were reported,
65.6 percent occurred at fixed facilities. (Exhibit F-l)
o We were able to identify industrial sectors for the
facilities and/or firms reported in association with
release events in 79 percent of the 829 events reporting
evacuations. (Exhibit 2.2-13)
2-18
-------
Events that occurred at distribution facilities (25.9
percent) and primary production facilities (25.0 percent)
accounted for the majority of evacuations. Events that
occurred at end users (14.5 percent) and secondary
producers (12.6 percent) accounted for the next largest
shares. Events occurring at firms involved with
waste-related activities accounted for only 1..1 percent
of the evacuation events. (Exhibit 2.2-13)
Compared to their frequency of involvement in all events,
distributors and end users have a slightly higher
frequency and primary producers a significantly lower
frequency of involvement in evacuation events. The rates
of involvement for facilities classified as secondary
producers and waste related are slightly lower for
evacuation events than for all events. (Exhibit 2.2-13)
For events that occurred at facilities coded as primary
producers, 4.8 percent involved evacuations. Similarly,
evacuations were reported for 6.8 percent of events at
secondary production facilities, 8.0 percent of the
events at distributors, and 9.7 percent of the events
reported at end-users. (Exhibit 2.2-14)
On average, more people were evacuated in distributor-
related events than in any other sector. Distributor-
related events accounted for 25.8 percent of the
evacuation events, and 38.0 percent of all persons
reported as evacuated. In contrast, end-user-related
evacuation events tend to involve a smaller number of
evacuees, i.e., while 14.5 percent of the evacuation
events involved end-users, these events contributed only
7.3 percent of all reported evacuees. (Appendix 0)
Of the 288 events in which evacuees can be identified as
workers or residents, residents, on average, accounted
for 90.7 percent of the persons evacuated while workers
accounted for 9.3 percent. The proportion of workers and
non-workers varied greatly by sector:
Primary Producers: 10.0 percent were workers
90.0 percent were non-workers
Secondary Producers: 34.5 percent were workers
65.5 percent were non-workers
Distributors: 2.9 percent were workers
97.1 percent were non-workers
2-19
-------
End-Users: 21.2 percent were workers
78.8 percent were non-workers
Waste Related: 82.9 percent were workers
17.1 percent were non-workers
2.2.4 Chemicals Involved
The ten chemicals most frequently reported in association
with evacuation events accounted for releases at up to
43 percent of such events.7 (Exhibit 2.2-16)
Chlorine was involved in the most events (89), that
representing 10.7 percent of evacuation events
reported in the AHE/DB. (Exhibit 2.2-16)
Hydrochloric acid, ammonia (anhydrous) and sulfuric
acid were involved in 76, 67 and 30 evacuation
events, respectively. Together, they accounted for
21 percent of such events. (Exhibit 2.2-16)
The six chemicals next most frequently reported in
conjunction with evacuation events accounted for 11
percent of such events. (Exhibit 2.2-16)
o Chlorine was the most frequently reported chemical in
association with evacuation events at primary producers,
secondary producers, end users and waste related
facilities. Chlorine was reported as released in 18
percent of the evacuation events at primary producers;
12 percent of the events at secondary producers; four
percent at distributors; 18 percent at end users; and 44
percent at waste disposal facilities. (Exhibits F-7,
F-9, F-ll, F-13 and F-15)
o After chlorine, the six substances next most frequently
reported as released at primary producers were
hydrochloric acid, ammonia (anhydrous), vinyl chloride,
ammonia, sulfuric acid, and nitric acid. Together, these
Since any given event may involve more than one chemical,
there is an overlap in the numbers of events, deaths and injuries
reported in association with more than one chemical. Therefore,
all percents or combined totals of events, deaths or injuries in
this section should be interpreted as upper bounds.
2-20
-------
chemicals account for 71 events, or 34 percent of the 207
events which occurred at primary production facilities.
(Exhibits F-6 and F-7)
For events that occurred at secondary producers, the
three chemicals most frequently reported as released, in
addition to chlorine, were ammonia (anhydrous), ammonia
and nitric acid. Together, these substances accounted
for 20 events, or 19 percent of the total number of
events that occurred at secondary producers. (Exhibit
F-9)
For events that occurred at distributors, the three most
frequently reported chemicals were hydrochloric acid,
ammonia (anhydrous) and sulfuric acid. Together they
accounted for 50 events or 23 percent of the total number
of events that occurred at distributors. (Exhibit F-ll)
The average quantities of substances released per event
for the 19 chemicals that were released during the
largest number of evacuation events are shown in Exhibit
2.2-16. Among evacuation events, releases of methyl
alcohol represent, on average, the largest quantities
(975,000 pounds of methyl alcohol were released during
evacuation events), followed by sulfuric acid (130,000
pounds) and phosphoric acid (100,000 pounds). However,
these average quantities are heavily influenced by a
small number of very large releases. Excluding the
largest single release for each substance changes the
rank ordering. The chemical with the largest adjusted
average release is still methyl alcohol (75,000 pounds
in one release), followed by styrene (15,000 pounds
average for five releases), hydrochloric acid (14,200
pounds average for 27 releases), sulfuric acid (11,800
pounds average for 12 releases), and ammonia (9,000
pounds average for eight releases).
When quantities of materials released in evacuation
events are compared to the reportable quantities (RQs)
for these substances, 25.4 percent of the evacuation
events for which release quantities were reported
involved quantities less than the RQ. That is, the
release of such quantities of substances would not be
reported to the National Response Center or most other
governmental jurisdictions. (Exhibit F-4)
Only 19.7 percent of the evacuation events involved the
release of substances in quantities greater than 100
times the released chemical's RQ. (Exhibit F-4)
2-21
-------
2.3 MOST SEVERE EVENTS INVOLVING ATMOSPHERIC RELEASES
OF ACUTELY TOXIC OR EXPLOSIVE CHEMICALS
As stated at the beginning of this chapter, some caveats are
necessary in interpreting this data. These are summarized below.
First, only a few of the reported events have been
independently verified. Thus, one should interpret the data as
"deaths or injuries reported by one or more of the contributing
sources." Second, direct evidence of causality rarely exists in
the contributing sources to the data base. That is, some of the
death or injuries reported to have occurred in association with a
particular event may not, in fact, have been caused by the release
of a chemical during an event. For example, an event involving the
release of a substance following a motor vehicle collision may
involve a death related to the impact of the collision. Although
deaths explicitly resulting from a collision were removed from the
data base whenever possible, it is not possible to identify
causality in all cases. Third, the term "injury" is used to refer
to physical manifestations ranging from temporary respiratory or
eye irritation treated on-site to critical injury leading to
prolonged hospital treatment. Finally, in some cases the source
of data used did not provide exact numbers of persons injured or
evacuated. In these cases standard rules were used to code such
terms as "a few" and "about 50" into specific numbers.
2.3.1 General Characteristics
o Air releases of toxic chemicals that led to two or more
deaths are listed in Exhibit 2.3-1. Events involving the
release of a chemical vapor, a fire or an explosion, or
a substance with a gaseous physical state are classified
here as "air releases." Thirty-six such events were
identified as having occurred during the five-year period
from 1982 to 1986, an average of seven events per year.
These events were reported to have led to 127 deaths in
total. These air release events comprised 32.1 percent
of all events with fatalities and 44.1 percent of the
total number of fatalities.
o The most serious single event, with respect to the number
of fatalities was an explosion and fire associated with
gunpowder and fireworks that led to 21 deaths in June of
1985.
2-22
-------
Four-fifths of these events reporting two or more
fatalities were limited to"two or three deaths, and only
two (five percent) of the events resulted in ten or more
deaths.
Events involving 40 or more injuries are listed in
Exhibit 2.3-2. Forty-two such events were identified as
having occurred during the five year period from 1982 to
1986, an average of eight events per year. These events
resulted in 4,121 injuries. These most severe air injury
events account for 4.3 percent of all injury events and
38.1 percent of all injuries.
The most serious single event, with respect to the number
of injuries, was a vapor release of aldicarb oxime in
August, 1985, which resulted in 430 injuries.
Three-quarters of the most severe events involving
injuries had 90 or fewer reported injuries, while only
four events (ten percent) resulted in 200 or more
injuries.
Events that involved 2,000 or more evacuees are listed
in Exhibit 2.3-3. Thirty-six such events were reported
during the five year period from 1982 to 1986,
representing, on average, seven events per year. These
events resulted in a total of 188,075 reported evacuees.
These most severe air evacuation events account for 4.3
percent of all evacuation events and 60.9 percent of all
evacuations.
The most serious event, with respect to the number of
evacuees was an explosion and fire associated with the
release of 22,500 pounds of phosphorus, causing the
evacuation of 40,000 people in July of 1986.
Two-thirds of the most severe evacuation events involved
4,000 or fewer evacuees, while three events (eight
percent) led to 10,000 or more evacuees.
Transportation events played a significant role in the
most severe evacuation events, accounting for one-third
(12) of the 36 most serious evacuation events.
Furthermore, transportation events account for 13 percent
of the events with deaths and 38 percent of the events
involving injuries. (Exhibit E-19) The most severe
transportation event with fatalities involved seven
deaths.
2-23
-------
o The number of most severe events involving deaths,
injuries and evacuations, shown in Exhibits 2.3-1 to
2.3-3, appears to have remained relatively constant over
the 1982 to 1986 period, with the exception of 1985.
This increase in 1985 may have been a result of greater
public interest and more comprehensive reporting of such
events in the immediate aftermath of the Bhopal tragedy
rather than an actual increase in the occurrence of
accidents. The number of severe incidents listed in
Exhibits 2.3-1 to 2.3-3 is too few and the time frame too
short to provide a reliable basis for evaluating these
types of trends.
2.3.2 Significance of Event End Effects
o The most significant factor in determining the severity
of events involving fatalities appears to be whether or
not a fire or an explosion occurred. Most of the events
in which two or more deaths occurred (listed in Exhibit
2.3-1) involved a fire and/or an explosion (86 percent) .
o Five (or 14 percent) of the events listed in Exhibit
2.3-1 occurred in situations where chemical toxicity
appeared to be the primary factor in determining event
severity, i.e., where a spill or vapor release occurred
but fire and explosion did not. In all cases, these
toxicity-dominated events were restricted to two or fewer
deaths.
o Toxicity appears to be a somewhat more important factor
in determining the severity of events involving injuries.
For injury events, toxicity was indicated to be the
primary factor in 25 (or 60 percent) of the most severe
events, and accounted for three of the four events which
resulted in 200 or more injuries.
o Fires and/or explosions appear to have had the most
significant role in determining the severity of
evacuation events, accounting for 25 (or 70 percent) of
the 36 evacuation events listed in Exhibit 2.3-3, the
three most severe events, each of which led to reported
evacuations of 10,000 or more persons.
2.3.3 Large Scale Domestic Releases of Extremely Toxic Chemicals
The potential toxic effects of a chemical release event vary
with the toxicity and the volume of substance released. Five
thousand two hundred and seventy eight events in the AHE/DB have
information on both quantities and acute toxicity. For those
2-24
-------
events, we divided quantity released by the IDLH value attributed
to that chemical to establish a relative measure of the potential
toxic effects of each of these releases. Exhibit 2.3-4 lists the
25 events in the AHE/DB having highest quantity-to-toxicity ratios.
A number of caveats are necessary in order to interpret
properly the information presented on Exhibit 2.3-4. First,
substances are commonly distributed in solution, the concentration
of which is dependent on the intended use of the product. Second,
since the IDLH value presented on Exhibit 2.3-4 is based on the
toxicity of a concentrated form of each substance, the ratio of
quantity released to toxicity for events involving the release of
a chemical in solution will be overstated. In addition, these
results are highly dependent on the accuracy of the information
gathered in developing the AHE/DB. Errors which overestimate the
quantity of a substance released will result in an over-estimate
of the ratio of quantity to toxicity. Given these caveats, the
reader is cautioned not to over-interpret the significance of this
analysis.
In addition, a number of substances appearing on Exhibit 2.3-
4 are not generally classified as highly toxic substances, and are
not comparable, in terms of toxicity, with many of the more
hazardous substances found in the AHE/DB. For example, sodium
hydroxide (caustic soda) appears on this list four times, possibly
due to the ubiquity of this substance in commerce and industry
(i.e., it is likely to be involved in more events simply due to its
common use), and to the large quantities of caustic soda commonly
stored and shipped (i.e., when released, the average quantity
released will be large). However, sodium hydroxide is not
generally considered an extremely hazardous substance.
The listing presented on Exhibit 2.3-4 is important not so
much for identification of particular events, or substances
released during these events, as for the insights it may provide
into those events with potentially severe end effects. Thus, it
can serve as a starting point in the identification of events for
which potentially catastrophic outcomes were avoided, due to event-
specific conditions, preparedness planning or intervention on the
part of plant personnel or emergency response personnel.
o By this measure, 25 events involved releases whose
quantity-to-toxicity ratio was greater than one-half that
generally associated with the release at Bhopal, and 17
events involved releases whose quantity-to-toxicity ratio
exceeded that of Bhopal.
o Among the 25 events whose quantity to toxicity ratio was
at least one-half of that associated with Bhopal (Exhibit
2.3-4), only one event reported deaths (two fatalities).
Five events reported injuries (ranging from five to 650
2-25
-------
people injured), and six reported evacuations (up to
20,000 people). This is a somewhat greater frequency of
severe consequences than that observed among all events.
The average numbers of injuries and evacuations were
somewhat greater for these 25 high-damage potential
events when compared with the AHE/DB as a whole. The
average number of deaths was approximately the same.
However, given the vast amount of materials released in
the events in these high-damage potential events, their
severity remains surprisingly low.
o The difference between the potential for human harm and
the actual consequences of these events appears to be
due, in part, to a combination of one or more of the
following factors:
The chemical's physical state at the time of release
was liquid or solid, and much of the material
remained liquid or solid when released, or other
physical/chemical properties prevented widespread
airborne exposure.
Few people were close to the point of release.
Weather conditions were favorable.
Preparedness at the site of release included
containment of materials released, or other
equipment or procedures to mitigate the extent of
release and population exposure.
Evacuation or other emergency response may have
removed potentially exposed victims.
2.3.4 Comparison with Major Industrial Disasters of Similar
Origin
o Ten of this century's largest industrial disasters
related to large scale atmospheric releases of toxic or
explosive chemicals are listed in Exhibit 2.3-5. Bhopal
is by far the most severe, accounting for more than 3,000
deaths to date, with reports of more than 200,000
injuries, 40,000 of which have been classified as
serious.
2-26
-------
The most serious event of this type to occur in the U.S.
is the explosion of a fertilizer ship in Texas City,
Texas in 1947. The total number of fatalities (552)
associated with this ammonium nitrate related event
places it in a category of only three other events (in
addition to Bhopal) where the death toll exceeded 500.
Toxicity represented the primary human health risk in
only two of these industrial disasters. One of these was
the 1976 release of TCDD in Seveso, Italy that caused 700
homes to be permanently abandoned and hundreds of animals
to die, but did not involve any human fatalities. Thus,
Bhopal is the only large scale chemical release event for
which substance toxicity can be attributed as the cause
of a large number of human deaths.
The average number of fatalities for the ten events
listed in Exhibit 2.3-5 is nearly 600, far larger than
the number of fatalities that occurred at any of the
events listed in Exhibit 2.3-1 for the 1982 to 1986
period. The largest number of fatalities among the
latter is 21.
An important factor contributing to the total number of
deaths in Bhopal was the lack of warning and instructions
in proper response given to the populace prior to and
during the period of release. In Seveso, Italy hundreds
of animals were killed because they could not be
evacuated effectively. The long term effect of the event
was the contamination of the soil and the abandonment of
a large land area.
Thus, the data in Exhibit 2.3-5 indicate that events of
severe magnitude have to date been associated either with
situations in which massive explosions have occurred or
in which toxic materials have been allowed to reach the
surrounding populace without a timely or effective
response by emergency response officials.
2-27
-------
3.0 CHARACTERISTICS OF ALL EVENTS
3.1 EXTENT OF DAMAGE
Injuries were reported for 9.1 percent of the 10,933
events in the AHE/DB. (Exhibit 2.2-1)
Deaths were reported for only 1.0 percent of the events.
(Exhibit 2.2-2)
Evacuations were reported for only 7.6 percent of the
events; however, for 46 percent of the events it was not
stated whether evacuation was required. (Exhibit D-l)1
A total of approximately one-half million people were
reported as evacuated as a result of chemical release
events. This total excludes non-quantitative
descriptions of number of people evacuated, such as
"several people" or "all employees." (Exhibit D-2)
In 288 events, accounting for 308,822 evacuees, it was
possible to distinguish between the evacuation of
residents and workers. These evacuation events involved
a total of over 280,000 residents and 28,000 workers.
(Exhibit D-2)
When evacuations were reported, an average of 1,076
people per event were evacuated. The number of people
evacuated ranged from one to 40,000 people. (Exhibit D-
2)
Property damage was reported for 13.5 percent of all
events. In 43.2 percent of the events property damage
was stated not to have occurred, and in 43.3 percent of
the events it was not stated whether such damage had
1As noted elsewhere in this report, exhibits using an
alphabetic designation, such as "D-2", refer to exhibits presented
in the appendices.
3-1
-------
occurred. Estimates of property damage are usually
expressed in very general terms, ranging up-to $100
million dollars for one event. (Exhibit D-l)
Similarly, less than three percent of the 10,933 events
were reported to have resulted in environmental damage,
while one-third were reported not to have resulted in
environmental damage. For the majority of events (64.2
percent), the reporting source does not state whether
environmental damage occurred. (Exhibit D-l)
3.2 CIRCUMSTANCES OF RELEASE
o Of the 10,933 events, 71 percent occurred at fixed
locations; the remainder occurred during transit.
(Exhibit 2.2-3)
o Spills were reported as an end effect in 68.3 percent of
the events; vapor release in 39.3 percent; fire in 6.7
percent; and explosion in 2.9 percent. The severity of
end effects was estimated to be (in decreasing order)
explosion, fire, vapor release and spill. Note that
multiple end-effects may be reported, therefore, percents
do not total to 100. (Exhibit 2.2-4)
o Of the 7,807 events for which quantity released was
reported, those events involving the release of more than
100,000 pounds of material (which represent three percent
of the events in the data base) accounted for 88.0
percent of the total quantity of reported releases.
(Exhibits 3.2-1 and D-6)
3.2.1 In-Plant Events
o Storage vessels, process vessels, and valves or piping
accounted for approximately equal shares of event
locations for the 7,760 in-plant events (22 percent to
28 percent each). (Exhibit 2.1-3)
o Events occurring at storage vessels accounted for 75.2
percent of the quantity of material released as a result
of in-plant events. However, when deaths or injuries are
involved, "vehicles not in-transit" accounted for 50.7
percent of the quantity released, while waste disposal
and pollution control equipment accounted for another
35.4 percent. In each case, the disproportionate shares
of quantities of materials released are attributable to
a single event with a large release quantity. (Exhibit
3.2-2)
3-2
-------
Equipment failure was identified as the cause of 46.9
percent of all events that occurred in-plant, but
accounted for only 32.0 percent of the quantity reported
as released in-plant. In comparison, operator error was
the attributed cause for 18.2 percent of the in-plant
events, while accounting for 46.0 percent of the quantity
released. For D/I events, sparking and lightning
accounted for 2.2 percent of the events, but over 50
percent of the quantity of material released. The large
share of material released attributable to sparking and
lightning was due to a single large quantity release
event. (Exhibits 2.1-4 and 3.2-3)
In-plant events primarily occurred under steady state
conditions (51.0 percent) or during loading or unloading
operations (16.2 percent). Operational context was not
reported for one-quarter of the incidents. (Exhibit 2.1-
5)
Less than six percent of the in-plant events accounted
for over 92 percent of the quantity of material released.
Those events that released more than 100,000 pounds of
substances (3.2 percent of in-plant events) accounted for
over 89 percent of the total quantity of material
released. (Exhibits 3.2-1 and 0-6)
Events that involved releases of over 100,000 pounds of
material accounted for the majority of material released
for all in-plant locations except valves/pipes,
heating/cooling and "other" locations. (Exhibits 3.2-4
through 3.2-9)
The majority of events that listed event cause as
equipment failure, upset conditions, fire/explosion, high
temperature and pressure, power related, sparking and
lightning, other or unknown involved the release of less
than 1,000 pounds of substances. The majority of events
that listed operator error, arson and vandalism or
disposal related activities as event cause involved the
release of less than 10,000 pounds of material.
(Exhibits 3.2-10 through 3.2-18)
Events that involved the release of over 100,000 pounds
of material accounted for the majority of material
released for all causes except disposal and power
related. (Exhibits 3.2-10 through 3.2-18)
3-3
-------
o Over 80 percent of the material released in all in-plant
event contexts (except the "other" category) was related
to events that released 50,000 pounds or more of
material. (Exhibits 3.2-19 through 3.2-23)
3.2.2 In-Transit Events
o In-transit events accounted for 29.0 percent of the
events in the AHE/DB. (Exhibit 2.1-7)
o Slightly less than one-half of all in-transit events
(48.6 percent) involved trucks; another 43.3 percent
involved rail cars. (Exhibit 2.1-8)
o For events involving the release of hazardous materials
during transportation, 46.3 percent were the result of
a leak, 11.6 percent were the result of a collision, 6.9
percent were the result of over-turned trucks and 6.5
percent were the result of derailments. (Exhibit 2.1-9)
o Over one-half of the events for each transit mode type
(except pipelines) involved releases of less than 1,000
pounds of material. The majority of events that occurred
from pipelines involved the release of over 1,000 pounds
of material. (Exhibits 3.2-24 through 3.2-28)
o The quantity reported as released from events occurring
at pipelines accounted for approximately one-half of the
total quantity of material released in in-transit events.
However, the bulk of this volume released was contributed
by a few large pipeline spills; otherwise, pipelines were
the reported transportation mode for only a small
fraction of release events and quantities released. The
quantities reported as released in rail and truck events
accounted for 22.3 and 18.5 percent of the total quantity
reported as released in in-transit events, respectively.
(Exhibit 3.2-29)
o Events that involved releases of over 100,000 pounds of
material accounted for almost all of the quantity
reported as released during pipeline and barge events.
For rail events, incidents involving the release of over
50,000 pounds of material accounted for over 85 percent
of the material released. (Exhibits 3.2-24 through 3.2-
28)
3-4
-------
Over 70 percent of the in-transit events that involved
rail transport involved releases of less than 1,000
pounds of material, over one-half released 100 pounds or
less of material, and 23.5 percent involved the release
10 pounds or less of material. (Exhibit 3.2-25)
No event that involved "other" transit modes (usually
air transport) reported a release quantity of over 25,000
pounds. (Exhibit 3.2-28)
For all in-transit event causes, except over-turned
trucks, the majority of material was released during
events that involved the release of over 100,000 pounds
of material. For events in which an "over-turned truck"
was identified as the cause of a release, approximately
50 percent of the total volume of the material was
released in volumes ranging from 1,000 to 50,000 pounds.
(Exhibits 3.2-30 through 3.2-33)
3.3 TYPES OF FACILITIES INVOLVED
o Facility type was identified for about 91 percent of the
events in the AHE/DB. Assigned facility types range from
primary producers to end-users, with 39.6 percent of the
events occurring in the Chemicals and Allied Products,
and Petroleum Products sectors (referred to in the AHE/DB
as the "Primary Producers" sector). Over 24 percent of
the events occurred in the distribution sector, 13.9
percent in the secondary producer sector, 11.3 percent
in end-user applications and 1.4 percent in waste-related
activities. In 9.2 percent of the events insufficient
information was provided in the reporting source to
identify facility type. (Exhibit 2.1-10)
o Although most events involving distributors are in-
transit events (69.8 percent), some of these events
occurred at fixed facilities. Similarly, while a large
majority of events involving primary producers (88.0
percent), secondary producers (94.0 percent), end users
(86.0 percent), and waste-related facilities (81.8
percent) occur at fixed facilities, a small fraction are
coded as in-transit events.
o Events occurring at primary production facilities
accounted for a greater percentage of reported deaths and
injuries than for all other industry sectors. Events
occurring at facilities classified as "distributors"
accounted for the largest share of evacuees. (Exhibits
2.1-11, 2.1-12)
3-5
-------
While events occurring at secondary production facilities
accounted for only 14 percent of all events, they
represent over 40 percent of the total reported quantity
of material released in the AHE/DB. Distributors and
end-users, on the other hand, represent twice as large
a percentage of the number of AHE/DB events as they do
of total quantity released. (Exhibits D-l, D-10, D-13,
D-16)
3.4 CHEMICALS INVOLVED
The ten chemicals most frequently reported in association
with AHE/DB accidental release events were involved in
43 percent of all reported events. (Exhibits 3.4-1 and
3.4-2)2
Releases of sulfuric acid were reported for the most
events: 1,045 or 9.6 percent of all of the events
in the AHE/DB. Five deaths and 365 injuries
occurred as a result of these events. (Exhibits
3.4-1 and 3.4-2)
Chlorine, ammonia (anhydrous) and hydrochloric acid
were involved, respectively, in 750, 628 and 530
events, together representing 17 percent of all
events in the AHE/DB. Seven deaths and 2,213
injuries occurred as a result of these events.
(Exhibits 3.4-1 and 3.4-2)
Six other chemicals (sodium hydroxide, methyl
chloride, phosphoric acid, ethylene oxide, toluene
and vinyl chloride) were involved in 16 percent of
all events reported in the AHE/DB. Twenty deaths
and 683 injuries occurred as a result of these
events. (Exhibits 3.4-1 and 3.4-2)
Sulfuric acid was the most frequently reported chemical
associated with events at all facility types except
"waste related." Sulfuric acid was involved in 11
percent of the events at primary producers, 11 percent
of the events at secondary producers and 9 percent of
the events at distributors. (Exhibits D-9, D-12 and D-
15)
2Since any given event may involve the release of multiple
chemicals, double counting exists when summing the number of
events, deaths or injuries reported for various chemicals.
Therefore, all percentages or combined totals of events, deaths or
injuries should be viewed as upper bound estimates.
3-6
-------
o After sulfuric acid, the next four most frequently
reported chemicals at primary producers were chlorine,
ammonia (anhydrous), methyl chloride and ethylene oxide.
These chemicals accounted for 1,195 events, or 28 percent
of the 4,324 events that occurred at primary producers.
(Exhibits D-8 and D-9)
o For events that occurred at secondary production
facilities, the five most frequently reported chemicals
were sulfuric acid, chlorine, ammonia (anhydrous),
phosphoric acid and hydrochloric acid. Together these
substances accounted for 509 events, or 33 percent of the
1,522 events that occurred at secondary production
facilities. (Exhibit D-12)
o For events that occurred at distributors, the four most
frequently reported chemicals, in addition to sulfuric
acid, were sodium hydroxide, hydrochloric acid, ammonia
(anhydrous) and phosphoric acid. Together they accounted
for 614 events, or 23 percent of the 2,689 events which
occurred at distributors. (Exhibit D-15)
o The average quantities of substances released per event
are shown in Exhibit 3.4-2 for the 20 chemicals
associated with the largest number of events. Releases
of sulfuric acid involved the largest quantities (88,000
pounds) , followed by sodium hydroxide (83,000 pounds) and
methyl alcohol (52,000 pounds).
o When release quantity is compared to the reportable
quantities (RQs) for the chemicals involved, 30.9 percent
of the events involved the release of substances in
quantities less than the RQ, i.e., these releases would
not normally be reported to the National Response Center
or most other jurisdictions. (Exhibit D-5)
o Only 13.1 percent of the events involved the release of
substances in quantities greater than 100 times their
RQs. (Exhibit D-5)
3.5 GEOGRAPHIC PATTERNS
Geographic distributions of all AHE events, death events,
injury events and evacuation events are shown in Exhibits 3.5-1
through 3.5-4. It is important to note three caveats regarding the
spatial patterns shown in these exhibits:
3-7
-------
The existence of high numbers of incidents in some states
may be a result of effective enforcement^of reporting
requirements by authorities in those statesrj.ower numbers
of incidents in other states may be a result of less
aggressive enforcement.
The distribution of events by state also reflects the
geographic emphasis of the data collection process. The
data collection process attempted to minimize such
biases, but was restricted by the need to access more
readily available data sources, such as those that had
been computerized. Exhibit 3.5-5 graphically depicts the
coverage achieved in each state. California, Texas, New
York and New Jersey received the most scrutiny.
We also would expect the geographic distribution of
events to be dependent upon the distribution of
manufacturing activity across states. That is, states
with extensive manufacturing activity, and thus, chemical
production and use, are likely to have higher numbers of
incidents. Exhibit 3.5-6 demonstrates that California,
Texas, New York, Illinois and Michigan all demonstrated
total values of shipments for manufactured products that
exceeded 120 billion dollars in 1985.
Exhibits 3.5-1 to 3.5-4 demonstrate the following:
o Texas was the only state with more than 1,000 events.
Ohio, New Jersey, New York, California and Louisiana also
demonstrated a large number of incidents. (Exhibit 3.5-
1)
o New York, New Jersey and Texas had the largest number of
release events involving deaths. (Exhibit 3.5-2)
o New Jersey and California had the largest number of
incidents involving injuries, followed by Texas,
Louisiana, Illinois, Ohio and New York. (Exhibit 3.5-3)
o The spatial distribution of evacuation events was similar
to that of injury events. New Jersey leads in the number
of events involving evacuations, followed by California
and Louisiana. (Exhibit 3.5-4)
3-8
-------
For each of the spatial distributions, the majority of
states fall in the categories representing the lowest
frequency of events: 39 states had 250 or fewer total
events; 44 states had four or fewer death events; 41
states had 25 or fewer injury events and 43 states had
25 or fewer evacuation events.
3-9
-------
Report Exhibits
-------
Exhibit 1.2-1
DATA SOURCES FOR THE AHE DATA BASE
National Data Sources
National Response Center
United Press International
Dept. of Transportation Hazardous
Material Information System
New York Times
PEI Associates
Associated Press
Marsh & McLennan
Chemical Week
Regional Data Sources
Region 7 EPA
New Orleans Tlmes-Plcayunne
Chicago Tribune
Los Angeles Times
Boston Globe
State Data Sources
Number of
Records
7,705
236
168
90
80
43
46
2
8.367
504
102
62
35
12
715
Period
1982-87
1980-85
1986-87
1965-87
1981-86
1980-86
1956-84
1985-86
1979-85
1980-86
1980-86
1980-86
1985-87
New Jersey Department of
Environmental Protection
New York Attorney General
Texas Air Control Board
Ohio EPA
California:
Office of Emergency Services
Highway Patrol
Additional Local Newspapers
513
433
428
252
133
61
24
Other
1.844
4
1980-87
1982-85
1980-86
1981-85
1984-85
1983-85
1980-86
TOTAL
10,933 *
National Response Center (NRC) data included 7,705 accidental
releases of Comprehensive Environmental Response. Compensation
and Liability Act (CERCLA) chemicals during 1982 through 1986.
Of these, the 2,435 events designated as air releases or having
reports of deaths, injuries or evacuations were all included in
the AHE Data Base. One-tenth of the remaining 5,270 events
events were randomly selected for inclusion In the AHE Data Base.
As a result, the 2.962 NRC-origin records In the AHE Data Base
represent 7,705 events: and the 6,190 total records in the AHE
Data Base represent 10,933 incidents.
Source: Industrial Economics, Inc.
-------
Exhibit 1.2-2
DISTRIBUTION OF EVENTS BY SOURCE AND YEAR
SOURCE
National Response Center
EPA Region 7
United Press International
DOTs Haz. Mat'l Inf. System
PEI Associates
Associated Press
Marsh & McLennan
Chemical Week
Newspapers:
New Orleans Times Picayunne
New York Times
Chicago Tribune
Los Angeles Times
Boston Globe
Additional Local Newspapers
States:
New Jersey Dept. of
Environmental Protection
New York Attorney General
Texas Air Control Board
Ohio EPA
California Office of
Emergency Services
California Highway Patrol
Other
TOTAL
Before
1980 or
Unknown
15
1
2
0
0
0
33
0
0
5
0
0
0
3
0
0
0
0
0
0
0
1980
0
1
3
0
0
19
4
0
26
9
4
4
0
0
66
0
19
0
0
0
0
1981
0
99
32
0
1
9
1
0
30
13
5
8
0
2
75
0
64
1
0
0
3
1982
1,065
141
38
0
3
3
3
0
10
12
4
7
0
0
80
2
63
70
0
0
0
1983
1.532
106
71
0
3
8
3
0
13
3
9
2
0
2
90
91
56
58
0
32
0
1984
1.860
106
58
0
1
3
2
0
6
1
13
1
0
2
57
149
122
100
111
26
1
1985
1.499
50
32
0
30
0
0
1
5
34
13
6
1
14
102
191
103
23
22
3
0
1986 1987
1.730
0
0
167
42
1
0
1
12
12
14
7
3
1
42
0
1
0
0
0
0
4
0
0
1
0
0
0
0
0
1
0
0
8
0
1
0
0
0
0
0
0
Total
7.705
504
236
168
80
43
46
2
102
90
62
35
12
24
513
433
428
252
133
61
4
59
155 343 1,501 2,079 2,619 2,129 2,033
15 10.933
Source: Industrial Economics, Inc.
-------
Exhibit 1.2-3
AHE/DB's Three Interconnecting Data Bases
NATIONAL SOURCES OF EVENTS DATA
National Response Center New York Times
DOTs HMIS PEI Associates
Marsh & McLennan Chemical Week
United Press International
Associated Press
REGIONAL SOURCES OF EVENTS DATA
Region 7 EPA New Orleans Times - Pfcayunne Boston Globe
Los Angeles Times Chicago Times Other Local Papers
STATE SOURCES OF EVENTS DATA
New Jersey DEQ
New York Attorney Genera.
Texas Air Control Board
Ohio EPA
EVENTS
DATA BASE
CHEMICALS
DATA BASE
COMPANY-
FACILITIES
DATA BASE
t
t
EPA / NIOSH Sources
RTECS
NFPA
Event Reports
Dun and Bradstreet
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-1
Occurrence of Injuries
DEATH/INJURY EVENTS
No Injuries Unknown
1.9% 0.5%
Injuries
97.6%
Number of Events » 1,019
ALL EVENTS
Unknown
Injuries
9.1%
No Injuries
8Z7%
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-2
Occurrence of Deaths
DEATH/INJURY EVENTS
Unknown Deaths
5.4% n.2%
No Deaths
83.4%
Number of Events = 1,019
ALL EVENTS
Unknown Deaths
5.9% 1.0%
No Deaths
93.1%
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-3
In-Plant Events By Location
DEATH/INJURY EVENTS
Unknown
17.6%
Process Vessel
22.1%
Vehicle not In Transit 4.2%
Heating/Cooling 1.7%
Disposal 3.0%
Valve/Pipe
15.8%
Storage Vessel
29.9%
Number of Events » 638
ALL EVENTS
Unknown
9.1%
Other 3.5%
Vehicle not In Transit 2.8%
Heating/Cooling 1.9%
Disposal 6.8%
Process Vessel
22.4%
Valve/Pipe
25.4%
Storage Vessel
28.1%
Number of Events = 7,760
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-4
In-Plant Events By Cause
DEATH/INJURY EVENTS
Other 2.0%
Sparking/Lightning 2.2%
Power 1.1%
High Pressure/Temp 2.7%
Arson/Vandalism 2.2%
Fire/Explosion
19.1%
Unknown
15.1%
Equipment Failure
30.9%
Upset Conditions 5.6%
Number of Events = 638
Operator Error
19.1%
Other 7.5%
Sparking/Lightning 0.3%
Power 1.7%
High Pressure/Temp 2.0%
Disposal 0.4%
Arson/Vandalism 0.9%
Fire/Explosion 5.2%
Upset Conditions
4.8%
ALL EVENTS
Unknown
1Z2%
Equipment Failure
46.9%
Operator Error
18.2%
Number of Events = 7,760
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-5
In-Plant Events By Event Context
DEATH/INJURY EVENTS
Unknown 32.0%
Loading/Unloading
16.5%
Other 0.8%
Maintenance 6.3%
Shut Down 0.8%
Start-Up 19%
Number of Events = 638
Steady State
39.8%
ALL EVENTS
Unknown 25.0%
Loading/Unloading
16.2%
Other 1.8%
Maintenance 3.1%
Shut Down 1.2%
Start-Up 1.8%
Steady State
51.0%
Number of Events = 7,760
Source: Industrial Economics, Inc.
-------
(D
UJ
Exhibit 2.1-6
Frequency of Reported End Effects
ALL EVENTS, DEATH/INJURY EVENTS
AND EVACUATION EVENTS
Vapor Release
Fire
Explosion
Other
End Effect
All Events
Death/Injury
Evacuation
Note: Summation across end effects may exceed 100% as each event may exhibit more than one end effect
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-7
In-Plant vs. In-Transit Events
DEATH/INJURY EVENTS
In-Transit
37.4% /:•:
In-Plant
62.6%
Number of Events = 1,019
ALL EVENTS
In-Transit
29.0%
In-Plant
71.0%
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-8
In-Transit Events By Mode
DEATH/INJURY EVENTS
Barge 2.6%
PipeVne
2.4% 1.3% other
Number of Events = 381
ALL EVENTS
Barge 4.5%
Pipeline
2.8% 0.7% Other
Truck
66.1%
Truck
48.6%
Number of Events = 3,173
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-9
In-Transit Events By Cause
DEATH/INJURY EVENTS
Unknown
4.5%
Other 14.7%
Truck Turnover 5.5%
Derailment 6.0%
Collision 22.8%
Leak
46.5%
Number of Events - 381
ALL EVENTS
Unknown
5.6%
Other 23.1%
Truck Turnover 6.9%
Derailment 6.5%
Collision 11.6%
Number of Events -3,173
Leak
46.3%
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-10
Distribution of Events By Facility Type
DEATH/INJURY EVENTS
Unknown
15.4%
Waste Related
1.3%
End User
13.1%
Primary Producer
26.2%
Distributor 31.1%
Number of Events = 1,019
Secondary Producer
13.0%
ALL EVENTS
Unknown
9.2%
Waste Related 1.4%
End User
11.3%
Distributor
24.6%
Primary Producer
39.6%
13.9% Secondary Producer
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-11
Reported Number of Deaths By Facility Type
Unknown
20.5%
Waste Related
2.4%
End User
17.4%
Primary Producer
43.4%
Distributor
7.6%
Secondary Producer
8.7%
Number of Deaths = 288
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-12
Reported Number of Injuries By Facility Type
Unknown
16.5%
Waste Related 0.7%
End User
8.6%
Distributor x£
25.3%
Primary Producer
31.0%
Secondary Producer
18.0%
Number of Injuries = 10,803
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-13
Frequency of Events Reporting Deaths
By Facility Type
2.0%-
2 1.5% -
I
in
1.0%-
0.5% -
0.0%
Number of Events with Typed Facilities = 9,928
Total Number of Events = 10,933
Primary
Producer
Secondary
Producer
Distributor End User
Facility Type
Waste
Related
All Types
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-14
Frequency of Events Reporting Injuries
By Facility Type
14.0%
12.0% -
10.0% -
0.0%
Number of Events with Typed Facilities = 9,928
Total Number of Events = 10,933
Primary
Producer
Secondary
Producer
Distributor End User
Facility Type
Waste
Related
All Types
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-15
Frequency of Reported End Effects
By Facility Type for Deattylnjury Events
Primary
Producer
Secondary
Producer
Distributor
End User
Waste
Related
Facility Type
Spill
Vapor Release
Fire
Explosion
Other
Total
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-16
Distribution of Events by Chemical
for Death/lnjury Events
10 CHEMICALS REPRESENTING
THE LARGEST NUMBER OF EVENTS
150-1
Substance Name
Note: Up to three substances may be reported as released during any one event in the AHE/DB. Thus,
double counting will result from summing the number of release events across substances.
Source: Industrial Economics, Inc.
-------
RANK
Exhibit 2.1-17
SUBSTANCES MOST FREQUENTLY REPORTED IN DEATH/INJURY EVENTS *
Notes
SUBSTANCE NAME'
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
CHLORINE
AMMONIA (ANHYDROUS)
HYDROCHLORIC ACID
SULFURIC ACID
AMMONIA
SODIUM HYDROXIDE
NITRIC ACID
TOLUENE
STYRENE
PHOSPHORIC ACID
TOLUENE DIISOCYANATE
BENZENE
PHOSGENE
SULFUR DIOXIDE
HYDROFLUORIC ACID
ACRYLONITRILE
METHYLENE CHLORIDE
HYDROGEN SULFIDE
ACETIC ACID
METHYL ALCOHOL
METHYL ETHYL KETONE
121
66
64
62
26
26
18
16
14
12
12
11
10
10
9
9
9
9
8
8
a
»OF
DEATH
AND
INJURY
(0/1)
EVENTS
121
66
64
62
26
26
18
16
14
12
12
11
10
10
9
9
9
9
8
8
a
fOF
DEATH
EVENTS
4
1
1
4
• 2
1
4
7
1
0
2
2
2
0
2
1
1
2
0
3
4
#OF
DEATHS
5
1
1
5
6
1
21
13
0
0
3
16
2
0
3
1
2
3
0
9
6
#OF
INJURY
EVENTS
121
65
64
61
26
26
18
15
13
12
11
11 .
10
10
8
9
8
9
8
8
7
#OF
INJURIES
1,054
248
911
365
308
39
361
70
32
369
112
436
303
55
664
63
481
157
41
29
101
TOTAL
QUANTITY
RELEASED
FROMD/I
EVENTS
(U>«.)
47.028
414,346
296,650
1.690.977
57,633
159,094
149,266
27.563
52,601
2.166
16,146
7.800
6,800
1,277
270.562
900.000
4,116
38
19,536
104.250
90
AVERAGE
QUANTITY
RELEASED
FROM 0/1
EVENTS
(Lbr) "•
644
9.636
10,595
51,242
5,763
10,606
16.585
9.188
8,767
361
3.229
3.900
3.400
426
90,187
225.000
1,372
38
3,907
34,750
90
LARGEST
QUANTITY
RELEASED
(Lbs.)
10.000
160.000
45.375
1.500.000
48.750
120,000
135.000
22.500
45.000
1.875
15.000
7.500
5,000
1.125
270.000
400.000
2.053
38
17,000
55.500
90
Includes all events In which the substance appears as the primary, secondary or tertiary substance.
All substances are commodity chemicals.
Average quantities are computed using only those events that reported quantities.
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-18
CHEMICAL PROPERTY CHARACTERIZATION METRICS
Chemical Property Basis for Determination
Rating Scale
Acute Toxicity
(LRATE)
NIOSH IDLH (level of concern) 0 1 2 3 4
or CEPP Level of Concern >500 50-500 5-50 0.5-5 0-0.5
LOW HIGH
Flammability
(FRATE)
NFPA Rating
Flash Point
Boiling Point
Qualitative Information
0...
LOW
..4
HIGH
Reactivity
(RRATE)
NFPA Rating
Qualitative Information
0...
LOW
.1 2. 3.,
4
HIGH
Volatility
(VRATE)
Vapor Pressure measured
in mm Hg
0...
<1
LOW
.1.
1-100
2
>100
HIGH
Physical State
Solid/Liquid/Gas (as
reported for each event)
Processing Mode
(COMMSPEC)
lEc's assessment of whether
substance is Commodity (high
production volume) or
Specialty (low production
volume)
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-19
FREQUENCY DISTRIBUTION OF AHE/DB CHEMICALS
BY CHEMICAL PROPERTY RATINGS
LRATE
FRATE
RRATE
VRATE
COMMSPEC
0
1
2
3
4
U
Total
0
1
2
3
4
U
Total
0
1
2
3
4
U
Total
0
1
2
U
Total
C
S
U
0
4
37
23
42
116
222
54
29
23
41
33
42
222
73
18
23
8
2
98
222
35
42
39
106
222
128
59
35
Total 222
Death and Injury
Chemicals
Number Percent
0.0%
1.8%
16.7%
10.4%
18.9%
52.3%
100.0%
24.3%
13.1%
10.4%
18.5%
14.9%
18.9%
100.0%
32.9%
8.1%
10.4%
3.6%
0.9%
44.1%
100.0%
15.8%
18.9%
17.6%
47.7%
100.0%
57.7%
26.6%
15.8%
100.0%
LRATE
FRATE
RRATE
VRATE
Total
0
1
2
3
4
U
Total
COMMSPEC
Total
0
1
2
U
Total
C
S
U
All Events
Chemicals
Number Percent
0
1
2
3
4
U
0
14
66
45
66
460
0
2.2%
10.1%
6.9%
10.1%
70.7%
651
115
82
49
111
65
229
651
Total
651
70
77
69
435
651
267
229
155
651
100.0%
17.7%
12.6%
7.5%
17.1%
10.0%
35.2%
100.0%
0
1
2
3
4
U
147
42
46
20
6
390
22.6%
6.5%
7.1%
3.1%
0.9%
59.9%
100.0%
10.8%
11.8%
10.6%
66.8%
100.0%
41.0%
35.2%
23.8%
100.0%
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-20
FREQUENCY DISTRIBUTION OF AHE/DB EVENTS
BY CHEMICAL PROPERTY RATINGS OF PRIMARY SUBSTANCE RELEASED
LRATE
FRATE
RRATE
VRATE
0
1
2
3
4
U
Total
0
1
2
3
4
U
Total
COMMSPEC
Total
0
1
2
U
Total
C
S
U
Total
All Events
Number Percent
0
96
1820
782
4484
3751
10933
4408
1154
364
1686
1433
1888
10933
10933
2079
1860
1557
5437
10933
8367
732
1834
10933
0.0%
0.9%
16.6%
7.2%
41.0%
34.3%
100.0%
40.3%
10.6%
3.3%
15.4%
13.1%
17.3%
100.0%
0
1
2
3
4
U
4534
873
1622
281
61
3562
41.5%
8.0%
14.8%
2.6%
0.6%
32.6%
100.0%
19.0%
17.0%
14.2%
49.7%
100.0%
76.5%
6.7%
16.8%
100.0%
LRATE
FRATE
RRATE
VRATE
0
1
2
3
4
U
Total
Total
COMMSPEC
Total
0
1
2
U
Total
C
S
U
Total
Death and Injury
Events
Number Percent
0
11
128
51
452
377
1019
1019
1019
150
143
111
615
1019
695
80
244
1019
0.0%
1.1%
12.6%
5.0%
44.4%
37.0%
100.0%
0
1
2
3
4
U
371
149
46
123
86
244
36.4%
14.6%
4.5%
12.1%
8.4%
23.9%
100.0%
0
1
2
3
4
U
433
71
121
16
4
374
42.5%
7.0%
11.9%
1.6%
0.4%
36.7%
100.0%
14.70/0
14.0%
10.9%
60.4%
100.0%
68.2%
7.9%
23.9%
100.0%
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-21
Role of Fire/Explosion vs. Toxicity as
Apparent Cause of Deaths or Injuries
Fire/
Explosion
Toxicity
% of Injuries % of Deaths
Injuries or Deaths
Source: Industrial Economics, Inc.
-------
1000
Exhibit 2.1-22
Event Severity By Toxicity Rating
TOXICITY AS CAUSE OF DEATH OR INJURY
100 ^
'c*
10-
» i
• *
«
•
•
• •
• •
*
•
1 2 3
Toxicity Rating
Note: See Exhibit 2.1-18 for explanation of rating metrics.
Source: Industrial Economics, Inc.
-------
Exhibit 2.1-23
Event Severity By Flammability Rating
FIRE/EXPLOSION AS CAUSE OF DEATH OR INJURY
1000
100-
(0
.2
*!Z
"£*
CO
I
-------
Exhibit 2.1-24
Event Severity By Reactivity Rating
FIRE/EXPLOSION AS CAUSE OF DEATH OR INJURY
1000
100-
(0
O
+
CO
0>
Q
10-
Reactivity Rating
Note: See Exhibit 2.1-18 for explanation of rating metrics.
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-1
Occurrence of Injuries
EVACUATION EVENTS
Unknown
6.0%
No Injuries
54.2%
Number of Events = 829
ALL EVENTS
Unknown
8.2%
Injuries
9.1%
No Injuries
82.7%
Injuries
39.8%
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-2
Occurrence of Deaths
EVACUATION EVENTS
Unknown Deaths
6.6% 4.8%
No Deaths
88.5%
Number of Events = 829
ALL EVENTS
Unknown Deaths
5.9% 1.0%
No Deaths
93.1%
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-3
In-Plant vs. In-Transit Events
EVACUATION EVENTS
In-Transit
34.4% /:;:
In-Transit
29.0%
In-Plant
65.6%
Number of Events =* 829
ALL EVENTS
In-Plant
71.0%
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
I
LU
Exhibit 2.2-4
Frequency of Reported End Effects
ALL EVENTS AND EVACUATION EVENTS
All Events
Evacuation
Explosion
Other
End Effect
Note: Summation across end effects may exceed 100% as each event may exhibit more than one end effect
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-5
In-Plant Events By Location
EVACUATION EVENTS
Unknown
14.2%
Process Vessel
19.5%
Other
4.8%
Vehicle not in Transit
5.7%
Heating/Cooling 1.7%
Disposal 4.8%
Valve/Pipe Ni
14.3%
Storage Vessel
35.1%
Number of Events - 544
ALL EVENTS
Unknown
9.1%
Other 3.5%
Vehicle not in Transit 2.8%
Heating/Cooling 1.9%
Disposal 6.8%
Process Vessel
22.4%
Valve/Pipe
25.4%
Storage Vessel
28.1%
Number of Events = 7,760
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-6
In-Plant Events By Cause
EVACUATION EVENTS
Unknown
11.0%
Other 3.1%
Sparking/Lightning 1.5%
Power 1.3%
High Pressure/Temp 2.4%
Disposal 0.2%
Arson/Vandalism 3.3%
Equipment Failure
33.3%
Fire/Explosion
21.5%
Upset Conditions 4.0%
Operator Error
18.4%
Number of Events = 544
ALL EVENTS
Unknown
12.2%
Other 7.5%
Sparking/Lightning 0.3%
Power 1.7%
High Pressure/Temp 2.0%
Disposal 0.4%
Arson/Vandalism 0.9%
Fire/Explosion 5.2%
Upset Conditions
4.8%
Equipment Failure
46.9%
Operator Error
18.2%
Number of Events = 7,760
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-7
In-Plant Events By Event Context
EVACUATION EVENTS
Loading/Unloading
9.6%
Unknown 23.4%
Other 0.9%
Maintenance 5.2%
Shut Down 0.7%
Start-Up 4.4%
Steady State
55.9%
Number of Events » 544
ALL EVENTS
Unknown 25.0%
Loading/Unloading
16.2%
Other 1.8%
Maintenance 3.1%
Shut Down 1.2%
Start-Up 1.8%
Steady State
51.0%
Number of Events = 7,760
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-8
In-Plant Events By Location
AMOUNT RELEASED IN EVACUATION EVENTS
Other 0.8% Process Vessel 3.0M>
Storage Vessel 9.9%
Valve/Pipe
3.3%
Disposal 83.0%
Total Amount » 20.0 Million Pounds
AMOUNT RELEASED IN ALL EVENTS
Unknown
Vehicle not in Transit 7.1% 0<9% Process Vessel 5.2%
Heating/Cooling 0.1%
Disposal 5.9%
Valve/Pipe
5.5%
Storage Vessel
75.2%
Total Amount = 346.2 Million Pounds
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-9
In-Plant Events By Cause
AMOUNT RELEASED IN EVACUATION EVENTS
Unknown
Other 1.1% O-4** 7.3% Equipment Failure
4.4% Operator Error
Z8% Upset Conditions
Fire/Explosion
84.0%
Total Amount =• 20.0 Million Pounds
AMOUNT RELEASED IN ALL EVENTS
Other 5.3%
Unknown
1.8%
Sparking/Lightning 6.8%
Power 0.2%
High Pressure/Temp 0.9%
Arson/Vandalism 0.5% .
Fire/Explosion 5.7%
Upset Conditions 0.8%
Equipment Failure
32.0%
Operator Error
46.0%
Total Amount = 346.2 Million Pounds
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-10
In-Transit Events By Mode
EVACUATION EVENTS
Pipeline
1.1%
Barge 3.2% O-4** Other
Number of Events = 285
ALL EVENTS
Barge 4.5%
Plpeine
2.8% 0.7% Other
Truck
50.2%
Truck
48.6%
Number of Events = 3,173
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-11
In-Transit Events By Cause
EVACUATION EVENTS
Unknown
3.9%
Other 8.1%
Truck Turnover 6.7%
Derailment 24.9%
Collision 21.4%
Number of Events - 285
Other 23.1%
Truck Turnover 8.9%
ALL EVENTS
Unknown
5.8%
Derailment 8.5%
Collision 11.6%
Number of Events -3,173
Leak
46.3%
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-12
In-Transit Events By Mode
AMOUNT RELEASED IN EVACUATION EVENTS
Barge
16.5%
Total Amount a 9.3 MilDon Pounds
AMOUNT RELEASED IN ALL EVENTS
Pipeline
48.3%
10.9% Barge
Total Amount = 53.4 Million Pounds
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-13
Distribution of Events By Facility Type
EVACUATION EVENTS
Unknown
21.0%
Waste Related
1.1%
End User
14.5%
Primary Producer
25.0%
Distributor 25.9%
Number of Events = 829
Secondary Producer
12.6%
ALL EVENTS
Unknown
9.2%
Waste Related 1.4%
End User
11.3%
Distributor
24.6%
Primary Producer
39.6%
13.9% Secondary Producer
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-14
Frequency of Events Reporting Evacuations
By Facility Type
i
§
UJ
12.0%
10.0% -
8.0%-
6.0%-
4.0%-
2.0%-
0.0%
Number of Events with Typed Facilities = 9,928
Total Number of Events =• 10,933
Primary
Producer
Secondary
Producer
Distributor End User
Facility Type
Waste
Related
All Types
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-15
Percentage of Evacuees by Facility Type
Residents versus Workers
Number of Events - 220
Total Evacuated - 259,030
10% H
o%
Residents
Workers
Primary
Producer
Secondary
Producer
Distributor
End User
Waste
Related
All Types
Facility Type
Source: Industrial Economics, Inc.
-------
Exhibit 2.2-16
SUBSTANCES MOST FREQUENTLY REPORTED IN EVACUATION EVENTS
BANK SUBSTANCE NAME *
1 CHLORINE
2 HYDROCHLORIC ACID
3 AMMONIA (ANHYDROUS)
4 SULFURIC ACID
5 AMMONIA
6 NITRIC ACID
7 VINYL CHLORIDE
8 SODIUM HYDROXIDE
9 HYDROGEN SULFIDE
10 STYRENE
11 TOLUENE
12 PHOSPHORIC ACID
13 TOLUENE DIISOCYANATE
14 PHOSGENE
15 METHYL ALCOHOL
16 ETHYLENE OXIDE
17 SULFUR DIOXIDE
18 POLYVINYL CHLORIDE
19 BENZENE
»OF
EVACUATION
EVENTS "
89
76
67
30
21
20
14
12
12
12
10
10
10
9
9
9
8
8
8
#OF
EVACUATIONS
(ALL SUBS)
38.037
47.243
12.126
14,055
5.158
17.124
17,686
5.070
3.022
17.250
2,950
18.760
4.958
1.109
3.275
6.225
630
18.250
5.050
TOTAL
QUANTITY
RELEASED
(Lb«)
52,777
1,508.389
415,086
1,641.870
183.612
152.776
150,696
40.723
1.669
157,238
46.875
300,563
1,058
100
1.950.000
225.026
0
0
19.500
AVERAGE
QUANTITY
RELEASED
(Lb«) —
1,199
53.871
11,860
126,298
20.401
19.097
37.674
8.145
334
26.206
15.625
100,188
529
50
975.000
56.257
0
0
9,750
LARGEST
QUANTITY
RELEASED
(Lbt)
20,000
1,125.000
160.000
1.500.000
112.500
135.000
150.000
30.000
1.000
82.500
30.000
300.000
938
90
1.875.000
202.500
0
0
12,000
Notes
All substances are commodity chemicals.
This Includes all events In which the substance appears as a primary, secondary, or tertiary substance.
Average quantities are computed using only those events that reported quantities.
Source: Industrial Economics. Inc.
-------
Exhibit 2.3-1
AIR RELEASES OF TOXIC CHEMICALS LEADING TO TWO OR MORE DEATHS
(1982-1986)
DATE
04-Apr-«6
12-Jun-86
17-Jan-8e
OI-May-86
27-May-8e
OS-Jun-86
os-Jui-ae
25-Jun-85
20-May-85
20-Oec-85
18-Apr-85
26-Apr-85
29-May-8S
19-Jan-85
06-Feb-85
03-Mar-e5
04-Apr-85
18-Jul-85
05-Dac-85
: SUBSTANCES RELEASED
SUBSTANCE1 SUBSTANCE2
GUNPOWDER AND/OR FIREWORK
TOLUENE NITROETHANE
AMMONIUM DICHROMATE
ACET ALDEHYDE
POLYVINYL CHLORIDE TOLUENE DIISOCYANATE
BUTADIENE FORMALDEHYDE
TRICHLOROETHYLENE
GUNPOWDER AND/OR FIREWORK
GUNPOWDER AND/OR FIREWORK
XYLENE TOLUENE
TOLUENE
POTASSIUM SODIUM NITRATE
AMMONIA
SULFUR
NITROGLYCERINE
HYDROFLUORIC ACID
ACETYLENE
METHYL ALCOHOL VARNISH
HYDROGEN NAPHTHA
END EFFECTS
SUBSTANCES 1ST 2ND 3RD 4TH
EX
AMMONIA Fl
EX
Fl
METHYLENE CHLORIDE Fl
SP
SP
EX
Fl
METHYL ETHYL KETONE VR
EX
EX
EX
Fl
EX
SP
Fl
Fl
EX
Fl
EX
Fl
SP
EX
EX
VR
Fl
EX
Fl
Fl
UU
UU
EX
UU
VR
EX
EX
Fl
UU
SP
VR
EX
VR
Fl
UU
UU
UU
EX
UU
UU
UU
SP
UU
UU
UU
UU
VR
UU
UU
UU
VR
UU
VR
UU
UU
UU
UU
UU
UU
UU
VR
UU
UU
UU
UU
UU
NUMBER
OF
INJURIES
0
4
18
2
0
6
25
6
0
7
1
17
10
7
0
0
0
1
45
NUMBER
OF
DEATHS
8
3
2
2
2
2
2
21
e
3
3
3
3
2
2
2
2
2
2
NUMBER
OF
EVACUEES
0
0
0
700
0
2.000
0
0
0
0
0
0
0
0
0
0
0
100
0
-------
Exhibit 2.3-1
(Continued)
AIR RELEASES OF TOXIC CHEMICALS LEADING TO TWO OR MORE DEATHS
(1882-1964)
DATE
11-May-84
16-Aug-84
10-Aug-«4
22-Aug-M
13-Oec-M
27-May-83
04-Dec-83
25-Jan~S3
28-Aug-83
ie-Nov-83
03-Oct-83
14-Jun-«3
21-Jul-«3
18-May-82
02-Jun-«2
13-Jan-82
14-Jan-82
<^8TANCE8BajEA8f33
SUB8TANCE1 SUBSTANCE2
PLASTIC POLYSTYRENE
ACETONE NITRIC AGIO
COAL TAR LIGHT OIL
SPECIALTY OILS
SULFURIC ACID
EXPLOSIVES
PETROLEUM PRODUCTS INSECTICIDE
EXPLOSIVES
SODIUM NITRATE
EXPLOSIVES
METHYL BROMIDE MALATHION
METHYL ISOBUTYL KETONE 2-NITROPROPANE
METHANE CHLORINE
NITROGEN
NITRIC ACID
HYDROGEN SULFIDE
UNKNOWN
'-:< 1
SUBSTANCES 1ST
URETHANE Fl
EX
EX
Fl
EX
EX
ANTIFREEZE Fl
EX
EX
Fl
EX
VR
EX
VR
EX
VR
EX
ENDEf
2ND
UU
Fl
Fl
UU
Fl
Fl
EX
Fl
Fl
EX
Fl
UU
Fl
UU
Fl
UU
Fl
•fECT!
3RD
UU
VR
UU
UU
UU
UU
VR
UU
UU
UU
VR
UU
SP
UU
VR
UU
UU
if:
4TH
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
VR
UU
SP
UU
UU
NUMBER
OF
INJURIES
0
1
21
1
0
1
4
0
10
20
6
0
0
3
04
15
11
NUMBER
OF
DEATHS
3
3
2
2
2
11
6
4
4
2
2
2
2
2
2
2
2
NUMBER
OF
EVACUEES
0
100
0
0
0
0
1.800
0
0
173
0
0
0
0
6.000
0
550
FIVE YEAR TOTAL
318
127
10.223
End Effect Codes: SP - spill; VR - vapor release: Fl - fire; EX - explosion; UU - unknown
Note: Only a small percentage of the information contained in the event reports used to construct the AHE/DB have been independently verified.
either by EPA or its contractors. Thus, one should interpret these data as 'deaths or injuries that were reported by one or more
of the contributing sources.* In addition, direct causation is rarely provided in the contributing sources to the AHE/DB. See additional
cautions for interpretation in Chapter 2.
Source: Industrial Economics, Inc.
-------
Exhibit 2.3-2
'l •'•
AIR RELEASES OF TOXIC CHEMICALS LEADING TO FORTY OR MORE INJURIES
(1082 -1086)
DATE
09-Jul-86
27-Mar-ee
2S-Apr-86
21-Fab-ee
22-Jul-8a
ie-Jul-86
30-May-se
11-Aug-8S
2t-Jan-8S
21-Mar-«5
26-Jun-8S
20-Apr-«5
05-Sep-85
30-Sap-8S
16-Jul-85
26-Jul-eS
13-Apr-85
22-Nov-85
04-Aug-a5
05-Dec-85
13-Mar-85
27-Aug-85
06-Ocl-84
10-May-84
15-Nov-84
08-Jan-84
22-Jan-84
17-Jul-fl4
25-OCI-84
SUaSTANCES RELEASED
SUBSTANCE1 SUBSTANCES
PHOSPHORUS
HYDROCHLORIC ACID
UNKNOWN
CHLORINE
ETHYLENE GLYCOL
SULFURIC ACID
UNKNOWN
ALDICARB OXIME METHYLENE CHLORIDE
DIMETHOATE
POLYVINYL CHLORIDE
PARATHION PARAQUAT
METHYL ETHYL KETONE PHOSGENE
UNKNOWN
CARBON MONOXIDE
POLYVINYL CHLORIDE HYDROCHLORIC AGIO
CHLORINE
BENZENE SULFURIC ACID
SILVER CYANIDE
EXPLOSIVES
HYDROGEN NAPHTHA
ACRYLIC RESIN
URANIUM OXIDE
MALATHION
HYDROCARBONS BENZENE
METHYL ISOCYANATE
HYDROGEN SULFIDE AMMONIA
PHOSPHORUS OXYCHLORID
AMMONIA
CHLORINE
SUBSTANCES 1ST
EX
SP
Fl
SP
SP
SP
VR
VR
VR
VR
GASOLINE VR
HYDROGEN CYANIDE Fl
VR
VR
Fl
VR
URANIUM EX
EX
Fl
EX
VR
SP
VR
LACQUER EX
VH
SP
SP
EX
SP
END EFFECTS
2ND 3RD 4TH
Fl
VR
VR
VR
VR
VR
UU
UU
SP
Fl
Fl
VR
UU
UU
VR
UU
Fl
VR
EX
Fl
SP
VR
UU
Fl
UU
VR
VH
VR
UU
VR
UU
UU
UU
UU
UU
UU
UU
UU
UU
EX
UU
UU
UU
UU
UU
VR
UU
UU
VR
UU
UU
UU
VR
UU
UU
UU
UU
UU
SP
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
NUMBER NUMBER
OF OF
INJURIES DEATHS
400
115
00
71
65
50
40
430
200
143
134
80
72
67
66
65
61
46
45
45
40
40
161
125
111
78
72
60
48
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
2
0
1
0
1
0
0
0
0
0
NUMBER
OF
EVACUEES
40,000
0
0
740
0
0
5.000
3.100
0
8.250
2.000
74
260
0
10.000
0
0
200
4.500
0
0
0
0
0
600
1.200
0
0
0
-------
Exhibit 2.3-2
(Continued)
AIR RELEASES OF TOXIC CHEMICALS LEADING TO FORTY OR MORE INJURIES
(1902-1880)
DATE
SUBSTANCE1
8UBSTANCE2
SUBSTANCES
END EFFECTS
1ST 2ND 3RD 4TH
NUMBER NUMBER
OF OF
INJURIES DEATHS
NUMBER
OF
EVACUEES
30-Mar-84
SULFURIC ACID
SP VR UU UU
40
200
IS-Jun-AS
10-Aug-«3
04-Apr-83
25-Jul-83
22-Dec-«2
13-Sep-«2
12-Nov-«2
31-Oct-«2
02-Jun-82
02-Jun-82
13-OCI-82
22-Oct-«2
PARATHION
SODIUM HYDROSULFIT
NITRIC ACID
UNKNOWN
METHYLACRYLATE
SODIUM SULFIDE
ETHYLENE OXIDE
CHLORINE
NITRIC ACID
PHOSGENE
VINYL BENZENE
TRICHLORO ETHYLENE
SP
SP
SP
EX
VH
Ft
VR
Fl
UU
VR
Fl
EX
UU
UU
EX
VR
118
70
40
40
0
0
0
0
1.600
107
6.000
300
PARAQUAT
FLUOROCARBON
TOLUENE DIISOCYANATE
ORTHENE
SP
VR
SP
SP
EX
VR
SP
SP
VR
Fl
VR
VR
Fl
SP
VH
VR
UU
UU
UU
UU
VR
UU
UU
UU
UU
UU
UU
UU
SP
UU
UU
UU
355
100
77
70
04
00
61
40
0
0
0
1
2
1
0
0
400
2.000
1.000
0
6.000
0
350
300
FIVE YEAR TOTAL
4.121
02.141
End Effect Codes: SP - spill: VR - vapor release; Fl - fire; EX - explosion; UU - unknown
Note: Only a small percentage of the Information contained in the event reports used to construct the AHE/OB have been Independently verified.
either by the EPA or its contractors. Thus, one should interpret these data as 'deaths or injuries that were reported by one or more of
the contributing sources.' In addition, direct causation is rarely provided in the contributing sources to the AHE/OB. See additional
cautions for interpretation in Chapter 2.
Source: Industrial Economics. Inc.
-------
Exhibit 2 £-3
AIR RELEASES OF TOXIC CHEMICALS LEADING TO TWO THOUSAND OR MORE EVACUEES
(1982-1086)
EOATE
08-JUM6
06-Aug-86
30-May-se
08-Jun-86
16-Jul-45
21-Mar-85
22-Jun-aS
27-Nov-8S
04-Aug-85
09-Jun-86
24-M«y-86
13-NOV-85
11-Aug-S5
16-Mar-85
10-Aufl-45
22-Mar-85
26-Jun-«5
04-Sep-flS
14-May-«4
31-Dec-84
10-Apr-«4
01-Jun-84
12-May-83
04-Apr-83
26-Jan-83
05-Nov-83
leUBSTANCES RELEASED
SUBSTANCE1 SUBSTANCE2 SUBSTANCES
PHOSPHORUS
CHLORINE
UNKNOWN
BUTADIENE
POLYVINYL CHLORIDE
POLYVINYL CHLORIDE
METHYL BROMIDE
LINURON
EXPLOSIVES
VINYL CHLORIDE
HYDROCHLORIC ACID
BROMINE
ALDICARB OXIME
TOLUENE DIISOCYANATE
CHLORINE
OCTYL ALCOHOL
PARATHION
HYDROCHLORIC ACID
SULFURIC ACID
ETHYLENE OXIDE
METHYL ALCOHOL
HYDROCHLORIC ACID
PLASTIC
NITRIC ACID
SULFURIC ACID
AMMONIA (ANHYDROUS)
CALCIUM HYPOCHLORITE
FORMALDEHYDE
HYDROCHLORIC ACID
PHOSPHATE AMMONIUM NITRATE
CALCINED ALUMINA
HYDROGEN SULFURIC ACID
M ETHYLENE CHLORIDE
PARAQUAT GASOLINE
COPPER SULFATE AMMONIUM PERSULFATE
PLASTIC
1ST
EX
Fl
VR
SP
Fl
VR
Fl
Fl
Fl
EX
EX
SP
VR
EX
VR
SP
VR
SP
SP
SP
SP
SP
VR
SP
SP
Fl
END EFFECTS
2ND 3RD
Fl
VR
UU
EX
VR
Fl
VR
EX
EX
Fl
Fl
VR
UU
Fl
UU
Fl
Fl
VR
VR
VR
Fl
VR
Fl
VR
VR
UU
VR
UU
UU
Fl
UU
UU
UU
VR
UU
SP
SP
UU
UU
VR
UU
UU
EX
UU
UU
UU
UU
UU
UU
Fl
UU
UU
4TH
SP
UU
UU
VR
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
UU
EX
UU
UU
NUMBER
OF
INJURIES
400
34
40
6
se
143
12
13
45
0
21
1
430
26
3
0
134
0
21
0
0
4
0
40
10
1
NUMBER
OF
DEATHS
0
0
0
2
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NUMBER
OF
EVACUEES
40.000
8.800
6.000
2.000
10.000
8.250
7.600
6.026
4.600
4.000
3.800
3.600
3.100
3.000
3.000
2.000
2.000
2.000
3.500
2,500
2.000
2.000
6.400
6.000
3.000
2.000
-------
Exhibit 2.3-3
(Continued)
AIR RELEASES OF TOXIC CHEMICALS LEADING TO TWO THOUSAND OR MORE EVACUEES
(1082-1M6)
EDATE
ll-Oec-82
02-Jun-«2
06-May-82
26-Feb-«2
28-Sep-82
30-Nov-«2
U-Sep-82
10-Apr-82
21-Apr-«2
13-Sap-«2
SUBSTANCES RELEASED
SUBSTANCE! SUBSTANCE2
ACROLEIN
NITRIC ACID
FUMARIC ACID
PHOSPHORUS THICHLORI
VINYL CHLORIDE TETRAETHYL LEAD
STYRENE
CHLORINE HYDROGEN PEROXIDE
CHLORINE
CHLORINE
SODIUM SULFIDE PARAQUAT
SUBSTANCES
HYDROCHLORIC ACID
SODIUM HYDROXIDE
ORTHENE
1ST
EX
EX
EX
SP
Fl
EX
EX
VR
Fl
VR
END EFFECTS
2ND 3RD
Fl
Fl
R
VR
EX
Fl
Fl
UU
VR
Fl
VR
VR
VR
UU
SP
VR
UU
UU
UU
UU
4TH
UU
SP
UU
UU
VR
UU
UU
UU
UU
UU
NUMBER NUMBER
OF OF
INJURIES DEATHS
0
64
10
2
6
11
1
12
0
100
0
2
0
0
0
0
0
0
0
0
NUMBER
OF
EVACUEES
20.000
6.000
3.600
3.000
3.000
3.000
2.500
2.100
2.100
2.000
FIVE YEAR TOTAL
1.643
184.075
End Effect Codec SP - •pill: VR - vapor rotate; Fl - firs; EX - explosion; UU - unknown
Note: Only a small percentage ol the information contained in the event reports used to construct the AHE/DB have been Independently
verified, either by the EPA or its contractors. Thus, one should interpret these data as 'deaths or injuries that were reported by one
or more of the contributing sources.' In addition, direct causation is rarely provided in the contributing sources to the AHE/DB. See
additional cautions for interpretation in Chapter 2.
Source: Industrial Economics. Inc.
-------
Exhibit 2.3-4
EVENTS WITH HIGHEST QUANTITY-TO-TOXICITY RATIO
Number
Plant or
Date
12/11/82
11/12/85
10/30/87
10/09/85
09/14/85
11/25/85
07/31/81
08/13/84
12/01/82
10/28/84
08/25/85
01/05/83
12/24/81
06/26/85
08/11/83
State Transit
LA P
FL
TX
MO
TX
N/A
KS
Ml
NE
CA
CA
AZ
NE
CA
NJ
P
P
T
T
P
T
P
T
P
T
P
P
P
P
Physical
State
GS
LQ
LQ
GS
LQ
LQ
LQ
LQ
LQ
LQ
LQ
LQ
LQ
LQ
LQ
GS
GS
GS
LQ
LQ
LQ
Substance(s)
Acrolein
Sulfuric Acid *
Hydrofluoric Acid * *
Hydrochloric Acid *
Sulfuric Acid *
Sodium Hydroxide *
Ammonia (Anhydrous)
Ammonia *
Hydrazine
EVENT TOTAL
Trichloroethane
Hydrogen Peroxide *
Sodium Hydroxide '
Sodium Hydroxide '
Nalco 1370
EVENT TOTAL
Sodium Hydroxide. *
Parathion
Paraquat
Gasoline
EVENT TOTAL
Hydrochloric Acid *
Phosphoric Acid *
Phosphorus Oxychloride
EVENT TOTAL
Quantity
(Ibs) |
480,000
6.000.000
270.000
1.125.000
1.500.000
1,500.000
1.575.000
1.500,000
N/A
1.500.000
23.250.000
375.000
1.125.000
938.000
N/A
938,000
750.000
3.750
3,750
4,500
12,000
N/A
N/A
51,600
51,600
IDLH
[gm/m3)
0.010
0.320
0.020
0.150
0.320
0.330
0.350
0.350
0.100
5.430
0.100
0.330
0.330
N/A
0.330
0.240
0.002
N/A
0.150
N/A
0.030
End Effects Number
Quant/IDLH 1st 2nd
48,000,000 EX Fl
18,750.000 SP UU
13.500.000 SP VR
7,500.000 SP VR
4.700.000 SP UU
4.500.000 SP UU
4.500,000 SP VR
4,300.000 SP UU
N/A
4,300.000
4.300.000 SP UU
3,800.000 SP UU
3,400,000 SP UU
2.800.000 SP UU
N/A
2,800,000
2,300.000 SP UU
16.000 VR Fl
1.900.000
N/A
1.900,000
N/A VR UU
N/A
1,700.000
1 .700,000
3rd 4th
VR UU
UU UU
UU UU
UU UU
UU UU
UU UU
UU UU
UU UU
UU UU
UU UU
UU UU
UU UU
UU UU
EX UU
UU UU
Injured
0
0
650
0
6
0
0
0
0
0
0
0
0
134
0
Hospit-
ajized FataMties
0 0
0
140
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Number
Non-Worke
Evacuated Evacuees
20,000 20,000
0
3.000
0
300
0
0
0
0
0
0
0
0
2.000
80
0
3,000
0
0
0
0
0
i
0
0
0
0
0
2,000
0
-------
Exhibit 2.3-4 (continued)
EVENTS WITH HIGHEST QUANTITY-TO-TOXICITY RATIO
Plant or Physical
Date State Transit State Substance(s)
Number
Quantity IDLH End Effects Number Hospit- Number Non-Worker
(Ibs) (gm/m3) Quant/IDLH 1st 2nd 3rd 4th Injured alized Fatalities Evacuated Evacuees
05/24/76
01/22/84
11/03/86
06/08/86
04/10/85
02/13/85
11/02/84
04/16/84
05/13/85
LA
IL
ME
TX
TX
TX
Wl
IL
OH
P
P
P
T
P
P
P
P
P
LQ
LQ
LQ
LQ
LQ
LQ
LQ
GS
GS
LQ
LQ
GS
GS
GS
Ethylene Oxide
Propylene Oxide
Polyglycol Ether
EVENT TOTAL
Phosphorus Oxychloride
Potassium Cyanide
Butadiene
Formaldehyde
EVENT TOTAL
Sulfur Dioxide
Sulfur Dioxide
Nitric Acid *
Hydrochloric Acid '
Acrylonitrile
Hydrogen Cyanide
Acetonitrile
EVENT TOTAL
2,400.000
N/A
N/A
2,400.000
45,000
75.000
466,600
157,500
624,100
339.000
324.000
277,500
150.000
37,000
61.000
N/A
98.000
1.440
4.730
N/A
0.030
0.050
44.080
0.120
0.262
0.262
0.260
0.150
8.650
0.070
6.690
1.700.000 EX Fl
N/A
N/A
1 ,700.000
1.500.000 SP VR
1.500,000 SP UU
11.000 SP EX
1.300.000
1 .300.000
1.300.000 VR UU
1.200.000 VR UU
1.100.000 SP UU
1.000,000 SP UU
4.000 VR UU
900.000
N/A
900.000
UU UU
UU UU
UU UU
Fl VR
UU UU
UU UU
UU UU
UU UU
UU UU
0
72
0
5
0
0
0
0
0
0
12
0
1
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
2.000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
End Effect Codes: SP = Spill; VR = Vapor .Release; Fl = Fire; EX = Explosion; UU = Unknown
N/A = Information was not reported in the sources used to characterize these events.
Note: * These substances are commonly distributed in solutions of less than full concentration. Thus acute toxicity (as measured by IDLH) may be overstated.
• * The firm that owns the facility at which this event took place submitted additional information indicating that the quantity reported as released in the
AHE/DB is overstated. Thus the quantity-to-toxicity ratio is overstated, though this event would remain on this list even if the quantity released is modified.
Source: Industrial Economics. Inc.
-------
Exhibit 2.3-5
TEN MAJOR CHEMICAL-RELATED INDUSTRIAL DISASTERS IN THE TWENTIETH CENTURY
Number of Number of Number of
Year
1984
1984
1984
1978
1976
1974
1948
1947
194Z
1921
Location
Bhopal. India
Mexico City. Mexico
Sao Paulo. Brazil
Los Alfaques, Spain
Seveso, Italy
Flixborough, England
Luduigshafen, FRG
Texas City, USA
Tessenderloo. Belgiua
Oppau. Germany
Description of Accident
Methyl isocyanate leak at chemical plant
LPG explosion at processing plant
Gasoline explosion following pipeline rupture
Propylene explosion following leak fro* lorry
TCDD leak at chemical plant
Cyclohexane explosion at chemical plant
Dimethyl ether explosion in factory
Aomoniin nitrate explosion while loading ship
Ammonium nitrate explosion in chemical plant
Ammonium nitrate explosion in fertilizer factory
Deaths
>3,000
>550
508
216
0
28
24S
552
200
561
Injuries
200.000
4.248
NA
NA
NA
NA
3,800
300
1.000
1.900
Evacuees
200,000
350,000
3,000
NA
>958
3.000
NA
NA
NA
NA
Sources: Kleindorfer and Kunreuther, 1987; Kletz, 1985; Chemical Engineering News. 1988.
Note: Seveso and Flixborough are included in this list because of their significance
in affecting public opinion regarding the dangers associated with chemicals
processing and transportation.
-------
Exhibit 3.2-1
Distribution By Release Size
for All Events
100% -/
90%-
Number of Events = 7,807
Amount Released - 399.6 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.5-6
Distribution of Value of Shipments of
Manufactured Products (1985)
Value of 1985 Shipments in $ Billions
• More than $179 Billion
E~3 $120 Billion to $179 Billion
E3 $60 Billion to $119 Billion
D Less than $60 Billion
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-2
In-Plant Events By Location
AMOUNT RELEASED IN DEATH/INJURY EVENTS
Unknown
0.3% 6.1% Process Vessel
Vehicle not in Transit
50.7%
6.0% Storage Vessel
1.3% Valve/Pipe
Disposal
35.4%
Total Amount » 46.6 Milton Pounds
AMOUNT RELEASED IN ALL EVENTS
Unknown
Vehicle not in Transit 7.1% a9% 5.2% Process Vessel
Heating/Cooling 0.1%
Disposal 5.9%
Valve/Pipe
5.5%
Storage Vessel
75.2%
Total Amount = 346.2 Million Pounds
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-3
In-Plant Events By Cause
AMOUNT RELEASED IN DEATH/INJURY EVENTS
Unknown 0.2% Equipment Failure 5.9%
Operator Error 2.1%
Upset Conditions 0.3%
Sparking/Lightning 50.7%
Fire/Explosion
35.9%
Mgh Pressure/Temperature 4.6%
Total Amount = 46.6 Million Pounds
AMOUNT RELEASED IN ALL EVENTS
Other 5.3%
Unknown
1.8%
Sparking/Lightning 6.8%
Power 0.2%
High Pressure/Temp 0.9%
Arson/Vandalism 0.5%
Fire/Explosion 5.7%
Upset Conditions 0.8%
Equipment Failure
32.0%
Operator Error
46.0%
Total Amount = 346.2 Million Pounds
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-4
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING PROCESS VESSELS
80%-
70%-
50% H
40%-j
Number of Events = 1,389
Amount Released = 18.1 MilGon Pounds
H Number
1 Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-5
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING STORAGE VESSELS
100%-^
90%-
80%-
70%-
*"" 50% H
-5
Number of Events = 1,665
Amount Released = 260.1 Million Pounds
a Number
H Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
100% •
90%-
Exhibit 3.2-6
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING VALVES OR PIPES
Number of Events = 1,746
Amount Released = 19.0 Milfion Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-7
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING DISPOSAL
100%-/
Number of Events = 370
Amount Released = 20.5 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
^
Exhibit 3.2-8
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING HEATING OR COOLING
90%-
80%-
70%-
3 60%^
*~ 50% H
40%-
Number of Events » 94
Amount Released = 0.5 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
.8
Exhibit 3.2-9
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING VEHICLES NOT IN TRANSIT
90%-
80%-
70%-
50%-
Number of Events = 177
Amount Released = 24.1 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-10
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING EQUIPMENT FAILURE AS CAUSE
100% •
90%-
Number of Events = 3,026
Amount Released = 110.9 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-11
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING OPERATOR ERROR AS CAUSE
90%-
80%-
70%-
50%-
40%-
30%-
20%-
10%-
Number of Events = 1,168
Amount Released 3 159.1 MilPon Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-12
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING UPSET CONDITIONS AS CAUSE
100% •/
90%-
Number of Events = 289
Amount Released = 2.8 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-13
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING FIRE OR EXPLOSION AS CAUSE
Number of Events = 104
Amount Released = 19.5 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-14
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING ARSON OR VANDALISM AS CAUSE
Number of Events = 42
Amount Released = 1.7 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
4—•
Exhibit 3.2-15
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING DISPOSAL AS CAUSE
90%-
80%-
70%-
50%-
40%-
30%-
20%-
10%-
0%'
Number of Events = 17
Amount Released = 0.1 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-16
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING HIGH PRESSURE
OR TEMPERATURE AS CAUSE
100% •/
90%-
80%-
70%-
Number of Events = 133
Amount Released = 3.1 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-17
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING POWER FLUCTUATIONS AS CAUSE
Number of Events a 102
Amount Released = 0.7 MilOon Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-18
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING SPARKING OR LIGHTNING AS CAUSE
ioo%-f
90%-
80%-
70%-
40%-
30%-
20%-
10%-
0%-
Number of Events = 6
Amount Released = 23.6 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
I
90%-
80%-
70%-
40%-(
30%-
20%-
10%-
Exhibit 3.2-19
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING LOADING OR
UNLOADING AS EVENT CONTEXT
Number of Events = 1,127
Amount Released = 113.3 Million Pounds
Number
Quantity
MO 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
.3
Exhibit 3.2-20
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING STEADY STATE CONDITIONS
AS EVENT CONTEXT
90%-
80%-
70%-
50%-
40%-
Number of Events = 3,045
Amount Released = 206.2 MilKon Pounds
Number
H Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-21
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING START-UP AS EVENT CONTEXT
100% -r
90%-
80%-
70%-
Number of Events = 101
Amount Released = 6.1 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-2SK 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-22
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING SHUT DOWN AS EVENT CONTEXT
Number of Events = 76
Amount Released = 1.1 Million Pounds
Number
Quantity
MO 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-23
Distribution By Release Size
for In-Plant Locations
EVENTS INVOLVING MAINTENANCE AS EVENT CONTEXT
3
100%-
90%-
80%-
70%-
50% H
40%-
30%-
20%-
10%-
Number of Events = 171
Amount Released a 2.5 MilOon Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K SOK-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-24
100% •
90%-
80%-
70%-
50%-
40%-
30%-
20%-
Distribution By Release Size
for In-Transit Events
INVOLVING TRUCKS
Number of Events = 1,170
Amount Released = 9.9 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-25
Distribution By Release Size
for In-Transit Events
INVOLVING RAIL
Number of Events = 523
Amount Released = 11.9 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-26
Distribution By Release Size
for In-Transit Events
INVOLVING PIPELINE
100% -1
90%-
Number of Events = 64
Amount Released = 25.8 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-27
Distribution By Release Size
for In-Transit Events
INVOLVING BARGE
100% -i
90%-
Number of Events = 106
Amount Released = 5.8 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-28
Distribution By Release Size
for In-Transit Events
INVOLVING OTHER MODES
Number of Events = 10
Amount Released = 32,296 Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-29
In-Transit Events By Mode
AMOUNT RELEASED IN DEATH/INJURY EVENTS
Barge
7.7% 0.1%
Truck
28.2%
Total Amount * 5.9 Million Pounds
AMOUNT RELEASED IN ALL EVENTS
Truck
18.5%
RpeUne
48.3%
10.9% Barge
Total Amount = 53.4 Million Pounds
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-30
Distribution By Release Size
for In-Transit Events
INVOLVING LEAKS AS CAUSE
Number of Events = 813
Amount Released - 6.9 Million Pounds
Number
Quantity
1-10 11-100
101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-31
90%-
80%-
70%-
- 60% H
50% H
40% H
30%-
20%-
10%-
Distribution By Release Size
for In-Transit Events
INVOLVING COLLISIONS AS CAUSE
Number of Events = 250
Amount Released = 10.6 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-32
Distribution By Release Size
for In-Transit Events
INVOLVING DERAILMENT AS CAUSE
100% y
90%-
80%-
70%-
Number of Events = 62
Amount Released = 4.8 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 25K-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-33
Distribution By Release Size
for In-Transit Events
INVOLVING OVER-TURNED TRUCKS AS CAUSE
o
10(3%-i
90%-
80% -
70%-
50% H
40%-
30%-
20%-
Number of Events = 181
Amount Released = 1.6 Million Pounds
Number
Quantity
1-10 11-100 101-1K 1K-10K 10K-25K 2SK-50K 50K-100K > 100K
Release Size (Ibs)
Source: Industrial Economics, Inc.
-------
Exhibit 3.2-34
Occurrence of Evacuations
DEATH/INJURY EVENTS
Unknown
35.5%
Unknown
46.1%
Evacuations
32.8%
No Evacuations
31.7%
Number of Events = 1,019
ALL EVENTS
Evacuations
7.6%
No Evacuations
46.4%
Number of Events = 10,933
Source: Industrial Economics, Inc.
-------
Exhibit 3.4-1
Distribution of Events by Chemical for All Events
i
0)
.a
3
10 CHEMICALS REPRESENTING
THE LARGEST NUMBER OF EVENTS
1200-1
1000-
800-
Substance Name
Note: Up to three substances may be reported as released during any one event in the AHE/DB. Thus,
double counting will result from summing the number of release events across substances.
Source: Industrial Economics, Inc.
-------
Exhibit 3.4-2
SUBSTANCES MOST FREQUENTLY REPORTED IN ALL EVENTS *
RANK SUBSTANCE NAME"
1 SULFURIC AGIO
2 CHLORINE
3 AMMONIA (ANHYDROUS)
4 HYDROCHLORIC ACID
5 SODIUM HYDROXIDE
6 METHYL CHLORIDE
7 PHOSPHORIC ACID
6 ETHYLENE OXIDE
9 TOLUENE
10 VINYL CHLORIDE
11 METHYL ALCOHOL
12 NITRIC ACID
13 TETRACHLOROETHYLENE
14 STYRENE
15 AMMONIA
16 HYDROGEN SULFIDE
17 ACETONE
18 METHYLENE CHLORIDE
19 BENZENE
20 METHYL ETHYL KETONE
WEIGHTED
TOTAL*
1.045
750
628
530
488
330
254
215
215
202
189
185
145
144
131
125
119
114
98
86
MAIN
205
690
518
250
88
280
44
215
75
172
49
75
35
54
111
115
29
44
48
36
SAMPLE
84
6
11
28
40
5
21
0
14
3
14
11
11
9
2
1
9
7
5
5
fOF
DEATH
EVENTS
4
4
1
1
1
0
0
0
7
2
3
4
0
1
2
2
1
1
2
4
fOF
DEATHS
5
5
1
1
1
0
0
0
13
6
9
21
0
0
6
3
3
2
16
6
tOF
INJURY
EVENTS
61
121
65
64
26
2
12
4
15
7
8
18
3
13
26
9
5
8
11
7
• OF
INJURIES
365
1.054
248
911
39
31
369
108
70
66
29
361
11
32
308
157
9
481
436
101
• OF
EVACUATION
EVENTS
30
89
67
76
12
5
10
9
10
14
9
20
2
12
21
12
3
6
8
4
• OF
EVACUEES
14.055
38.037
12.126
47.243
5.070
4.600
18.760
6.225
2.950
17.686
3.275
17.124
25
17.250
5.158
3.022
451
3.750
5.050
474
TOTAL
QUANTITY
RELEASED
(Lbi)
70,097.694
182.907
3.569.711
5.416,514
30.454.552
187.419
2,807.019
2.698.023
2,137.992
250.110
7,321,262
4,110.549
828.663
520.660
1.977.861
1,167.942
1,056.318
141.309
1.777.925
333.650
AVERAGE
QUANTITY
RELEASED
(Lbs)++
87.513
343
8.187
14.719
82.533
689
15.173
14.584
18.920
1.573
51.924
32,623
5.962
5.985
31,395
13.904
14.671
1.766
23.394
4.766
LARGEST
QUANTITY
RELEASED
(Lb.)
6.000.000
20.000
1.575.000
1.125.000
1.500.000
37.500
300.000
2,400.000
172,500
150,000
1,875.000
277.500
67.500
82.500
1,500,000
1,000.000
63.000
7.500
400.000
30.000
Notes:
Includes all events in which the substance appears as the primary, secondary or tertiary substance.
All chemicals are commodity chemicals.
The weighted total equals MAIN plus ten times SAMPLE.
Average quantities are computed using only those events that reported quantities.
Source: Industrial Economics, Inc.
-------
Exhibit 3.5-1
Distribution of All Events
Number of Events
• More than 1000 Events
m 501 to 1000 Events
El 251 to 500 Events
H 101 to 250 Events
D 51 to 100 Events
D 10 to 50 Events
D Less than 10 Events
Source: Industrial Economics, Inc.
-------
Exhibit 3.5-2
Distribution of Death Events
Number of Death Events
• More than 10 Events
B 5 to 9 Events
ED 3 to 4 Events
EJ 2 Events
ED 1 Event
D No Events
Source: Industrial Economics, Inc.
-------
Exhibit 3.5-3
Distribution of Injury Events
Number of Injury Events
• More than 100 Events
• 51 to 100 Events
ED 26 to 50 Events
H 11 to 25 Events
Q 6 to 10 Events
D 1 to 5 Events
D No Events
Source: Industrial Economics, Inc.
-------
Exhibit 3.5-4
Distribution of Evacuation Events
Number of Evacuation Events
• More than 100 Events
m 51 to 100 Events
D 26 to 50 Events
Q 11 to 25 Events
E3 6 to 10 Events
D 1 to 5 Events
D No Events
Source: Industrial Economics, Inc.
-------
Exhibit 3.5-5
Distribution of AHE/DB Sources
Source Types
• National, Regional and State
E3 National and State Only
E3 National and Regional Only
D National Only
Source: Industrial Economics, Inc.
------- |