CHAPTER 4: OFFSITE CONSEQUENCE ANALYSIS

RMP OFFSITE CONSEQUENCE ANALYSIS GUIDANCE

This chapter is intended for people who plan to do their own air dispersion modeling. If you plan to
do your own modeling, this chapter will provide you with the basic information you need to comply
with the rule requirements; it does not provide modeling methodologies. For people who do not plan
to do their own modeling, EPA has prepared a separate document, RMP Offsite Consequence
Analysis Guidance, which provides simple methods and reference tables for determining distance to
an endpoint for worst-case and alternative release scenarios. In conjunction with the National
Oceanographic and Atmospheric Administration (NOAA), EPA has also developed a software
program, RMP*Comp™, that performs calculations described in the RMP Offsite Consequence
Analysis Guidance. RMP*Comp™ is available for free from the EPA Internet website at
[Link] In addition, EPA has
developed industry-specific guidance for several industries covered by part 68, which are appended
to this document. In the appendices, EPA provides chemical-specific modeling for the applicable
industries, using the methods described in the RMP Offsite Consequence Analysis Guidance.

4.1 INTRODUCTION

The offsite consequence analysis consists of two elements:

� Worst-case release scenario analysis to identify the potential reach and

effect of hypothetical worst-case accidental releases, as follows:

� For a process to qualify for Program 1 (see Chapter 2), a worst-case

release scenario analysis must be done for each toxic and flammable
substance held above the applicable threshold quantity in the
process. The process is eligible for Program 1 if there are no public
receptors within the distance to an endpoint for any of the worst-case
scenarios analyzed and the other Program 1 criteria are met (see
Chapter 2). Consequently, for every Program 1 process, you must
conduct a worst-case scenario analysis for each regulated substance
above the threshold in the process. However, in your RMP the only
worst-case scenario analysis you must report for a Program 1
process is for the toxic or flammable substance that, if released
under worst-case conditions, would potentially effect the largest
geographical area around your facility. In the language of the rule,
such a scenario results “in the greatest distance to an endpoint,”
meaning that the dangerous effect of the toxic or flammable
substance extends the longest distance before dissipating to
relatively harmless levels.

� If your site has Program 2 and/or Program 3 processes (processes

that are not eligible for Program 1 — see Chapter 2), you must
identify, analyze and report one worst-case scenario to represent all
of the toxic regulated substances held above threshold quantities in
Chapter 4
Offsite C onsequence A nalysis 4-2

Program 2 and/or 3 processes (i.e., the worst-case release of the

toxic substance that results in the greatest distance to endpoint,), and
one worst-case scenario to represent all of the flammable regulated
substances held above threshold quantities in Program 2 and/or 3
processes (i.e., the worst-case release of the flammable substance
that results in the greatest distance to endpoint).

� You may need to submit an additional worst-case analysis if a

worst-case release from elsewhere at your facility could potentially
affect public receptors different from those potentially affected by
the initial worst-case scenario(s).

� Alternative release scenario analysis to identify the potential reach and

effect of hypothetical accidental releases under more realistic circumstances,
for all Program 2 and Program 3 processes, as follows:

� You must identify and analyze reasonable release scenarios that are
more likely to occur than worst-case release scenarios and that reach
an off-site endpoint, unless no such scenario exists.

� In your RMP, you must report one alternative release scenario

analysis for each regulated toxic substance held above the applicable
threshold quantity in a Program 2 or 3 process, including the
substance considered in a worst-case analysis.

� In your RMP, you must also report one alternative release scenario
representing all flammable substances held above the applicable
threshold quantity in a Program 2 or 3 process.

If the distance to the endpoint for your worst-case release just reaches your fence
line, you may not have an alternative release scenario with a distance to an endpoint
that goes beyond the fence line. However, you still must report an alternative release
scenario.

HOW SHOULD I CO NDU CT THE ANALYS IS ?

As noted above, You may use EPA's RMP Offsite Consequence Analysis Guidance
to carry out your consequence analysis. Results obtained using the methods in EPA's
Guidance are expected to be conservative. Conservative assumptions have been
introduced to compensate for high levels of uncertainty. EPA's guidance is optional,
and you are free to use other air dispersion models, fire or explosion models, or
computation methods provided that:

� They are publicly or commercially available or are proprietary models that

you are willing to share with the implementing agency;

� They are recognized by industry as applicable to current practices;

� They are appropriate for the chemicals and conditions being modeled;
 Chapter 4
4-3 Offsite C onsequence A nalysis

� You use the applicable definitions of worst-case scenarios; and

� You use the applicable parameters specified in the rule.

Complex models that can account for many site-specific factors may give less
conservative estimates of offsite consequences than the simplified methods in EPA's
guidance, particularly for alternative scenarios, for which EPA has not specified

EXHIBIT 4-1
CONSIDERATIONS FOR CHOOSING A MODELING METHOD

Approach Examples Advantages Disadvantages

Simp le EP A’s Offsite � Free � Conserva tive results

guidance Consequence � No com puter requirements � Few site-specific factors
Analysis � Simple to use considered
Guidance � Provides all data needed � Little flexibility in scenario
� Provides tables of distances development
� Eases compliance with rule

Simp le EP A mo dels, � No/low cost � Some may not be simple to use

computer such as � May be simple to use � Likely to give conservative
mod els RMP*Comp™ � Can consider some site- results
specific factors � May not accep t all of EP A’s
required assumptions
� May not inc lude chemical-
specific data
� May not address all
consequences

Complex Commercially � May address a variety of � May be c ostly

computer availab le scenarios � May require high level of
mod els mod els � May consider many site- expertise
specific factors

Calculation “Yellow Boo k” � Low co st � May require expertise to apply

methods (Netherlands � No com puter requirements methods
TNO) � May require de velop ment of a
variety of data

many assumptions. However, complex models may be expensive and require

considerable expertise to use; EPA's optional guidance is designed to be simple and
straightforward. You will need to consider the tradeoff in deciding how to carry out
your required consequence analyses. Exhibit 4-1 provides additional suggestions on
making this decision.

Whether you use EPA's guidance or another modeling method, you should bear in
mind that the results you obtain from modeling your worst-case or alternative
Chapter 4
Offsite C onsequence A nalysis 4-4

scenarios should not be considered to predict the likely results of an accidental

release. The worst-case assumptions are very conservative, and, regardless of the
model used, you can expect very conservative results. Results from modeling
alternative scenarios will be less conservative; however, you still must use
conservative endpoints.

In addition, results of an actual release will depend on many site-specific conditions

(e.g., wind speed and other weather conditions) and factors related to the release
(e.g., when and how the release occurs, how long it takes to stop it). You should
make reasonable assumptions regarding such factors in developing your alternative
scenarios, but the circumstances surrounding an actual release may be different.
Different models likely will provide different results, even with the same
assumptions, and most models have not been verified with experimental data;
therefore, results of even sophisticated modeling have a high degree of uncertainty
and should be viewed as providing a basis for discussion among the regulated
community, emergency planners and responders, and the public, rather than
predictions. Modeling results should be considered particularly uncertain over long
distances (i.e., 10 kilometers or more).

Exhibit 4-2 provides suggestions for assistance on modeling.

WHEN DOES THE OFFSITE CONSEQUENCE ANALYSIS (OCA) NEED TO BE REVISED?

You’ll need to revise your OCA when a change at your facility results in the distance
to an endpoint from a worst-case release rising or falling by at least a factor of two.
For example, if you increase your inventory substantially or install passive
mitigation to limit the potential release rate, you should re-estimate the distance at an
endpoint. If the distance is at least doubled or halved, you must update and res-
submit the RMP. For most substances, the quantity that would be released would
have to increase (or decrease) by more than a factor of five to double (or halve) the
distance to an endpoint. Depending on the regulated substance, installation of
passive mitigation systems such as containment dikes or enclosures can significantly
affect the distance to an endpoint.

4.2 WORST-CASE RELEASE SCENARIOS

This section describes the assumptions you must make and what you need to do to
meet the requirements for worst-case scenario analysis under the rule. Exhibit 4-3
summarizes the required parameters for the worst-case analysis. EPA has defined a
worst-case release as the release of the largest quantity of a regulated substance from
a vessel or pipe that results in the greatest distance from the point of release to a
specified endpoint, beyond which serious injury is not expected to occur (§68.3).
You must estimate the distance as follows:

� Part 68, Appendix A lists the toxic endpoint you must use for each regulated
toxic substance. For the worst-case analysis for toxic substances, you are
required to estimate the air dispersion distance to the endpoint, using certain
conservative assumptions concerning quantity released and release
conditions.
 Chapter 4
4-5 Offsite C onsequence A nalysis

� A vapor cloud explosion is specified as the worst-case scenario for

flammable substances. For the worst-case analysis for flammable
substances, you need to estimate the distance to an overpressure endpoint of
1 pound per square inch (psi) resulting from a vapor cloud explosion of a
cloud containing the largest quantity of the regulated flammable substance
from a vessel or process pipe line failure.

EXHIBIT 4-2
POSSIBLE SOURCES OF ASSISTANCE ON MODELING

� You may be able to obtain modeling help from the implementing agency in your area; for example,
local and state implementing agencies in Delaware, Florida, Georgia, and other “delegated” states (see
Chapter 10) are prepared to provide assistance to regulated sources. If you are in another state, you
may obtain help from the EPA regional office for your state.

� If you use certain models, users’ groups may be a source of assistance; for example, there is an
ALOHA model users’ group.

� If you use a commercial model, you probably can request assistance from the model developer or
distributor.

� Publications of the Center for Process Safety of the American Institute of Chemical Engineers (AIChE)
may provide useful information on modeling; examples of such publications include:

< Guidelines for Evaluating the Characteristics of Vapor Cloud Explosions, Flash Fires, and
BLEVEs (1994), and

< Gu ideline s for U se of V apo r Cloud D ispersio n M ode ls (1987).

� EPA publications also may provide useful modeling information; examples include:

< Wo rkbo ok of Scree ning Tech niqu es for A ssessing Imp acts of Toxic Air P ollutants, EPA­
450/4-88-009 (September 198 8), and

< Guidance on the App lication of Re fined D ispersion M odels for H azardo us/Toxic A ir Release,
EPA -454/R-93-002 (M ay 1993).

< Technical Background Document for Offsite Consequence Analysis for Anhydrous Ammonia,
Aqueous Ammonia Chlorine, and Sulfur Dioxide EPA-550-B99-017.

< EP A guidance is available at the Sup port Ce nter for Regulato ry Air Models:
[Link] .[Link]/scram0 01//
Chapter 4
Offsite C onsequence A nalysis 4-6

WORST -CASE RELEASES OF TOXIC SUBSTANCES

For the worst-case release analysis for toxic substances, you need to use the
assumptions discussed below, the properties of the substance, and an appropriate air
dispersion model or EPA's optional guidance to estimate the distance from the
release point to the point at which the concentration of the substance in air is equal
to the toxic endpoint specified in the rule. Because the assumptions required for the
worst-case analysis are very conservative, the results likely will also be very
conservative. The endpoints specified for the regulated toxic substances are
intended to be protective of the general public. These endpoints are concentrations
below which it is believed nearly all individuals could be exposed for one-half to one
hour without any serious health effects. In addition, the worst-case analysis is
carried out using very conservative assumptions about weather and release
conditions. The distance to the endpoint estimated under worst-case conditions
should not be considered a zone in which the public would likely be in danger;
instead, it is intended to provide an estimate of the maximum possible area that
might be affected in the unlikely event of catastrophic conditions. Distances greater
than about 10 kilometers are particularly uncertain. EPA intends the estimated
distances to provide a basis for a discussion among the regulated community,
emergency planners and responders, and the public, rather than a basis for any
specific predictions or actions.

MODELING ASSUMPTIONS

Quantity. EPA has defined (§68.3) a worst-case release as the release of the largest
quantity of a regulated substance from a vessel or process line (pipe) failure that
results in the greatest distance to a specified endpoint. For substances in vessels, you
must assume release of the largest amount in a single vessel; for substances in pipes,
you must assume release of the largest amount in a pipe. The largest quantity should
be determined taking into account administrative controls. Administrative controls
are written procedures that limit the quantity of a substance that can be stored or
processed in a vessel or pipe at any one time, or, alternatively, occasionally allow a
vessel or pipe to store larger than usual quantities (e.g., during turnaround). You do
not need to consider the possible causes of the worst-case release or the probability
that such a release might occur; the release is simply assumed to take place.

Release Height. All releases are assumed to take place at ground level for the
worst-case analysis. This is a conservative assumption in most cases. Even if you
think a ground-level release is unlikely at your site, you must use this assumption for
the worst-case analysis.

Wind Speed and Atmospheric Stability. Meteorological conditions for the

worst-case scenario are defined in the rule as atmospheric stability class F (stable
atmosphere) and wind speed of 1.5 meters per second (3.4 miles per hour). If,
however, you can demonstrate that the minimum wind speed at your site (measured
at 10 meters) has been higher than 1.5 meters per second, or that the maximum
atmosphere stability has always been less stable than class F, you may use the
minimum wind speed and most stable atmospheric conditions at your site for the
worst-case analysis. To demonstrate higher minimum wind speeds or less stable
atmospheric conditions, you will need to document local meteorological data from
 Chapter 4
4-7 Offsite C onsequence A nalysis

the previous three years that are applicable to your site. If you do not keep weather
data for your site (most sources do not), you may call another nearby source, such as
an airport, or a compiler, such as the National Weather Service, to determine wind
speeds for your area. Exhibit 3-1 in Chapter 3 describes atmospheric stability classes
in relation to wind speed and cloud cover. Your airport or other source will be able
to give you information on cloud cover. A small difference in wind speed probably
will not lead to a significant decrease in the distance to the endpoint.

EXHIBIT 4-3
REQUIRED PARAMETERS FOR MODELING WORST-CASE SCENARIOS

Endpo ints
� For toxic substances, use the endpoint specified in part 68, Appendix A.
� For flammable substances, use the endpoint of an overpressure of 1 pound per square inch (psi) for
vapor cloud exp losions.

W ind speed/stability
� Use wind speed of 1.5 meters per second and F stability class unless you can demonstrate that local
meteorological data applicable to the site show a higher minimum wind speed or less stable atmosphere
at all times during the previous three years. If you can demonstrate a higher minimum wind speed or
less stable atmo sphere over three years, these minim ums m ay be used.

Amb ient temperature/humidity

� For toxic substances, use the highest daily maximum temperature during the past three years and
avera ge hum idity for the site.

Height of release
� For toxic substances, assume a ground level release.

Topography
� Use urban or rural topography, as appropriate.

Dense or neutrally buoyant gases

� Tables or models used for dispersion of regulated toxic substances must appropriately account for gas
density.

Temperature of released substance

� For liquids (other than gases liquefied by refrigeration), use the highest daily maximum temperature,
based on data for the previous three years, or a t proc ess temperature, whicheve r is higher.
� Assume gases liquefied by refrigeration at atmospheric pressure are released at their boiling points.

Temperature and Humidity. For the worst-case release of a regulated toxic

substance, you must assume the highest daily maximum temperature that occurred in
the previous three years (the highest temperature reached in the last three years) and
the average humidity for the site. If you have not kept information on temperature
and humidity at your site, you may obtain it from a local meteorological station.
EPA's RMP Offsite Consequence Analysis Guidance assumes a temperature of 25°C
(77°F) and 50 percent humidity. If you use the EPA's guidance for your offsite
consequence analysis, you may use these assumptions even if the actual highest
temperature at your site is higher or lower. If the temperature at your site is
significantly lower, EPA's guidance may give overly conservative results,
Chapter 4
Offsite C onsequence A nalysis 4-8

particularly for toxic liquids. Small differences in temperature and humidity are
unlikely to have a major effect on results, however.

Topography. Two choices are provided for topography for the worst-case scenario.
If your site is located in an area with few buildings or other obstructions, you should
assume open (rural) conditions. If your site is in an urban location, or is in an area
with many obstructions, you should assume urban conditions.

Gas or Vapor Density. For the worst-case analysis, you must use a model
appropriate for the density of the released gas or vapor. Generally, for a substance
that is lighter than air or has a density similar to that of air, you would use a model
for neutrally buoyant vapors. The initial vapor density of a substance with respect
to air can be estimated from its molecular weight, assuming air has a "molecular
weight" of approximately 29. For a substance that is heavier than air (molecular
weight greater than 29), you generally would use a dense gas model. There are cases
where a dense gas model may be appropriate for a substance with molecular weight
of 29 or less (e.g., release of a compressed gas as a cold vapor) or where a neutrally
buoyant plume model may be appropriate for a substance with a higher molecular
weight (e.g., release by slow evaporation, with considerable mixing with air). In
addition, dense gases and vapors will become neutrally buoyant through mixing with
air as they move downwind. If you can account for such conditions in modeling, you
may do so.

ESTIMATING RELEASE RATES

Toxic Gases. Toxic gases include all regulated toxic substances that are gases at
ambient temperature (temperature 25º C, 77º F). For the consequence analysis, the
total quantity in the single largest vessel or process line is assumed to be released as
a gas over a period of 10 minutes, except in the case of gases liquefied by
refrigeration under atmospheric pressure. The release rate (per minute) for a gas (not
liquefied by refrigeration) is the total quantity released divided by 10. Passive
mitigation measures (e.g., enclosure) may be taken into account in the analysis of the
worst-case scenario. A 10-minute release must be assumed for gases regardless of
the model you use.

Gases liquefied by refrigeration alone (not under pressure) and released into diked
areas may be modeled as liquids at their boiling points, if the pool formed by the
released liquid would be greater than one centimeter (0.39 inches) in depth. In this
case, you may assume the liquefied gas is released from a pool by evaporation at the
boiling point of the gas. If the refrigerated liquefied gas is not contained by passive
mitigation, or if the pool formed would have a depth of one centimeter or less, you
must treat the released substance as a gas released over 10 minutes. EPA's analysis
indicated that pools of gas liquefied by refrigeration with a depth of one centimeter
or less would evaporate so rapidly at their boiling points that treatment as gaseous
releases over 10 minutes is reasonable.

Toxic liquids. For toxic liquids, you must assume that the total quantity in a vessel
is spilled, forming a pool. For toxic liquids carried in pipelines, you must assume
that the largest quantity that might be released from the pipeline forms a pool.
Passive mitigation systems (e.g., dikes) may be taken into account in consequence
 Chapter 4
4-9 Offsite C onsequence A nalysis

analysis. You must assume that the total quantity spilled spreads instantaneously to
a depth of one centimeter (0.39 inches) in an undiked area or covers a diked area
instantaneously. You estimate the release rate to air as the rate of evaporation from
the pool. To estimate the evaporation rate, you need to estimate the surface area of
the pool. You can take into account the surface characteristics of the area into which
the liquid would be spilled; for example, some models for pool evaporation will take
into account the type of soil if the spill will take place in an unpaved area. Your
modeling also should consider the length of time it will take for the pool to
evaporate.

You may use any appropriate model to estimate the evaporation rate of a spilled
regulated substance from a pool and estimate the air dispersion distance to the
specified endpoint of the regulated substance. The release rate can then be used to
estimate the distance to the endpoint.

ESTIMATING DISTANCE TO THE ENDPOINT

You may use any appropriate model, as discussed above, to estimate the distance to
the endpoint specified in part 68 Appendix A for a release of a regulated toxic
substance, using the required modeling assumptions.

WORST -CASE RELEASES OF FLAMMABLE SUBSTANCES

For the worst-case scenario involving a release of a regulated flammable substance

(a flammable gas or volatile flammable liquid), you must assume that the quantity of
the flammable substance is released into a vapor cloud and that a vapor cloud
explosion results. You must estimate the distance to an endpoint to an overpressure
level of 1 pound per square inch (psi) from the explosion of the vapor cloud.

� If the flammable substance is normally a gas at ambient temperature and

handled as gas or liquid under pressure or if the flammable substance is a gas
handled as a refrigerated liquid and is not contained when released or the
contained pool is one centimeter or less deep, you must assume the total
quantity is released as a gas and is involved in a vapor cloud explosion.

� If the flammable substance is a liquid or a refrigerated gas released into a

containment area with a depth greater than one centimeter, you may assume
that the quantity that volatilizes in 10 minutes is involved in a vapor cloud
explosion.

As in the case of the worst-case release analysis for toxic substances, the worst-case
distance to the endpoint for flammable substances is based on a number of very
conservative assumptions. Release of the total quantity of a flammable substance in
a vessel or pipe into a vapor cloud generally would be highly unlikely. Vapor cloud
explosions are also unlikely events; in an actual release, the flammable gas or vapor
released to air might disperse without ignition, or it might burn instead of exploding,
with more limited consequences. The endpoint of 1 psi is intended to be
conservative and protective; it does not define a level at which severe injuries or
death would be commonly expected. An overpressure of 1 psi is unlikely to have
serious direct effects on people; this overpressure may cause property damage such
Chapter 4
Offsite C onsequence A nalysis 4-10

as partial demolition of houses, which can result in injuries to people, and shattering
of glass windows, which may cause skin laceration from flying glass.

To carry out the worst-case consequence analysis for flammable substances, you may
use a TNT-equivalent model (i.e., a model that estimates the explosive effects of a
flammable substance by comparison with the effects of an equivalent quantity of the
high explosive trinitrotoluene (TNT), based on the available combustion energy in
the vapor cloud - see Guidelines for Evaluating the Characteristics of Vapor Cloud
Explosions, Flash Fires, and BLEVEs (1994) for additional information). Such
models allow you to estimate the distance to a specific overpressure level, based on
empirical data from TNT explosions. If you use a TNT-equivalent model, you must
assume that 10 percent of the flammable vapor in the cloud participates in the
explosion (i.e., you assume a 10 percent yield factor for the explosion). You do not
have to use a TNT-equivalent model; other models are available that take into
account more site-specific factors (e.g., degree of confinement of the vapor cloud).
Generally, however, a TNT-equivalent model is the simplest to use.

NUMBE R OF SCENARIOS

The number of worst-case scenarios you must analyze depends on several factors as
discussed below. You only need to consider the hazard (toxicity or flammability) for
which a substance is regulated (i.e., even if a regulated toxic substance is also
flammable, you only need to consider toxicity in your analysis; even if a regulated
flammable substance is also toxic, you only need to consider flammability).

PR OG RA M 1 PROCESSES

To demonstrate that a process is eligible for Program 1 (see Chapter 2), you must
conduct a worst-case release analysis for every toxic and flammable regulated
substance held in that process above the applicable threshold quantities. For the
process to be in Program 1, that analysis must show that the distance to the specified
endpoint for every regulated substance in the process is smaller than the distance to
any public receptor. If you have several processes that may qualify for Program 1,
you will have to conduct worst-case analysis for each process to determine which
qualify. You will need to report in the RMP the worst-case results with the greatest
distance to an endpoint for those processes you determine to be eligible for Program
1.

If the distance to the endpoint in the worst-case analysis is equal to or greater than
the distance to any public receptor, the process would be in Program 2 or Program 3
(discussed below). When you consider possible eligibility of your processes for
Program 1, you may want to look particularly at processes containing only
flammable substances, which are likely to have shorter worst-case distances than
toxic substances.

PR OG RA M 2 AND 3 PROCESSES

For all your Program 2 and 3 processes taken together, you must identify, carry out
and report in the RMP one worst-case analysis for the regulated toxic substances and
one worst-case analysis for the regulated flammable substances held above their
 Chapter 4
4-11 Offsite C onsequence A nalysis

threshold quantities in those processes. The basic purpose of the worst-case analysis
is to identify the geographical areas and the public receptors and population within
those areas that could be affected by a worst-case release. The release that results in
the greatest distance to an endpoint would affect the largest geographical area and
the greatest number of public receptors and people, so as a general matter only that
release (and not others affecting a smaller area) needs to be reported. The reported
scenario for toxic substances must be the scenario estimated to result in the greatest
distance to a toxic endpoint; for flammable substances, it must be the scenario
estimated to lead to the greatest distance to 1 psi overpressure for a vapor cloud
explosion. Additional worst-case analyses must be reported for toxic or flammable
substances if a worst-case release from a different location at the facility potentially
would affect a somewhat different geographical area containing public receptors and
people in addition to those affected by the scenario giving the greatest distance.

IDENTIFYING THE "WORST " WORST -CASE SCEN ARIO

Toxics. To determine the scenario that gives the greatest distance to an endpoint for
processes containing toxic substances, you may have to analyze more than one
scenario, because the distances depend on more than simply the quantity in a
process. For toxic liquids, for example, distances depend on the magnitude of the
toxic endpoint, the molecular weight and volatility of the substance, and the
temperature of the substance in the process, as well as quantity. A smaller quantity
of a substance at an elevated temperature may give a greater distance to the endpoint
than a larger quantity of the same substance at ambient temperature. In some cases,
it may be difficult to predict which substance and process will give the greatest
worst-case distance. You also may need to carry out analyses of worst-case
scenarios for locations at significant distances from each other to determine whether
different public receptors might be affected by releases.

Flammables. For flammable substances, the greatest quantity in a process is likely

to give the greatest distance to the endpoint, but there may be variations, depending
on heat of combustion and distance to the fence line. You may have to carry out
several analyses to identify the scenario that gives the greatest distance to the
endpoint. As in the case of toxic substances, you also may need to carry out analyses
for locations far apart from each other to determine whether different geographical
areas and thus public receptors might be affected.

For both toxic and flammable substances, the worst-case distances should be
considered only approximations.
Chapter 4
Offsite C onsequence A nalysis 4-12

QS & AS
WORST -CASE AND MITIGATION

Q. At my facility, if the worst-case release scenarios for regulated toxic substances and the worst-
case scenario for regulated flammable substances involve the same process, must I analyze both?

A. Yes. If the worst-case release scenarios for regulated toxic substances and regulated flammable
substances in Program 2 and 3 processes are associated with the same process, the two worst-case
release scenarios must be analyzed separately.

Q. What measures qualify as "passive mitigation"?

A. Passive mitigation is defined in § 68.3 as "equipment, devices, or technologies that function

without human, mechanical, or other energy input." Passive mitigation systems include building
enclosures, dikes, and containment walls. Measures such as fire sprinkler systems, water curtains,
valves, scrubbers, or flares would not be considered passive mitigation because they require human,
mechanical, or energy input to function.

Q. When analyzing the worst-case scenario for regulated toxic substances, must I anticipate a
specific cause (e.g., fire, explosion, etc.) of the scenario?

A. No. The worst-case analysis for a release of regulated toxic substances must conform to specific
assumptions as identified in § 68.25(c) and (d). Anticipated causes of the release will not affect the
analysis, and are not required. A specific cause may be considered in analyzing the alternative
release scenarios although it is not a requirement.

Q. Would all of the regulated substances stored in a salt dome be assumed to be released in the
worst-case scenario?

A. The worst case scenario for salt domes would be examined in a manner similar to that for
underground storage tanks. Reservoirs or vessels sufficiently buried underground are passively
mitigated or prevented from failing catastrophically. You should evaluate the failure of piping
connected to underground storage for the worst-case and alternative scenarios.

Q. Are valves in piping considered administrative controls?

A. No, administrative controls are written procedures that limit the quantity stored or flowing
through the pipes. Valves are considered active mitigation systems.

4.3 ALTERNATIVE RELEASE SCENARIOS

There are only a few required assumptions for the alternative scenario analysis.
Exhibit 4-4 summarizes the required assumptions.
 Chapter 4
4-13 Offsite C onsequence A nalysis

EXHIBIT 4-4
REQUIRED PARAMETERS FOR MODELING ALTERNATIVE SCENARIOS

Endpoints
� For toxic substances, use the endpoints specified in part 68, Appendix A.
� For flammable substances, use as the endpoints:
< Overpressure of 1 pound per square inch (psi) for vapor cloud explosions,
< Radiant heat of 5 kilowatts per square meter (kW/m2 ) (or equivalent dose) for
fireballs or pool fires, or
< Lower flammability limit (LFL) for vapor cloud fires.

Wind speed/stability
� Use typical meteorological conditions at your site.

Ambient temperature/humidity
� Use average temperature/humidity data gathered at your site or at a local meteorological
station.

Height of release
� Release height may be determined by the release scenario.

Topography
� Use urban or rural topography, as appropriate.

Dense or neutrally buoyant gases

� Tables or models used for dispersion of regulated toxic substances must appropriately
account for gas density.

Temperature of released substance

� Substances may be considered to be released at a process or ambient temperature that is
appropriate for the scenario.

ACCEPTABLE ALTERNATIVE SCENARIOS

Your alternative scenario for a covered Program 2 or 3 process must be one that is
more likely to occur than the worst-case scenario and that reaches an endpoint
offsite, unless no such scenario exists. You do not need to demonstrate greater
likelihood of occurrence or carry out any analysis of probability of occurrence; you
only need to use reasonable judgement and knowledge of the process. If, using a
combination of reasonable assumptions, modeling of a release of a regulated
substance from a process shows that the relevant endpoint is not reached offsite, you
can use the modeling results to demonstrate that a scenario does not exist for the
process that will give an endpoint offsite. Even in that instance, you still must report
an alternative scenario (i.e., one with an onsite endpoint) for Program 2 and 3
processes. Alternative scenarios are not required for Program 1 processes.
Chapter 4
Offsite C onsequence A nalysis 4-14

Release scenarios you should consider include, but are not limited to, the following,
where applicable:

� Transfer hose releases due to splits or sudden uncoupling;

� Process piping releases from failures at flanges, joints, welds, valves and
valve seals, and drains or bleeds;

� Process vessel or pump releases due to cracks, seal failure, drain bleed, or
plug failure;

� Vessel overfilling and spill, or overpressurization and venting through relief

valves or rupture disks; and

� Shipping container mishandling and breakage or puncturing leading to a

spill.

For alternative release scenarios, you may consider active mitigation systems, such
as interlocks, shutdown systems, pressure relieving devices, flares, emergency
isolation systems, and fire water and deluge systems, as well as passive mitigation
systems. Mitigation systems considered must be capable of withstanding the event
that triggers the release while remaining functional.

You should consider your five-year accident history and failure scenarios identified
in your hazard review or process hazards analysis in selecting alternative release
scenarios for regulated toxic or flammable substances (e.g., you might choose an
actual event from your accident history as the basis of your scenario). You also may
consider any other reasonable scenarios.

The alternative scenarios you choose to analyze should be scenarios that you
consider possible at your site. Although EPA requires no explanation of your choice
of scenario, you should choose a scenario that you think you can explain to
emergency responders and the public as a reasonable alternative to the worst-case
scenario. For example, you could pick a scenario based on an actual event, or you
could choose a scenario that you worry about, because circumstances at your site
make it a possibility. If you believe that there is no reasonable scenario that could
lead to offsite consequences, you may use a scenario that has no offsite impacts for
your alternative analysis. You should be prepared to explain your choice of such a
scenario to the public (as well as to the implementing agency), should questions
arise.

ALTERNATIVE RELEASES OF TOXIC SUBSTANCES

To estimate distances to the endpoint for alternative releases of toxic substances, you
need to identify reasonable scenarios for the regulated substances in covered
processes at your site and model these scenarios using appropriate models. As noted
above, for alternative release scenarios, you are permitted to take credit for both
passive and active mitigation systems, if both are in place. Modeling alternative
releases of toxic substances is discussed below.
 Chapter 4
4-15 Offsite C onsequence A nalysis

Although alternative scenarios are intended to be more likely than worst-case

scenarios, the analysis of alternative scenarios should not be expected to provide
realistic estimates of areas in which the public might be endangered in case of a
release. The same conservative, protective endpoints are used for alternative release
analysis as for worst-case analysis. These endpoints are intended to represent
exposure levels below which most members of the public will not suffer any serious
health effects. The endpoints are based on exposures for longer periods than may be
likely in an actual release. In addition, modeling carried out to estimate distances to
these endpoints, even when based on more realistic assumptions than used for the
worst-case modeling, likely will provide results with a high degree of uncertainty.
These estimated distances should not be considered a necessarily accurate prediction
of the results of an actual release.

MODELING ASSUMPTIONS

Quantity. EPA has not specified any assumptions you must make concerning
quantity released for an alternative release scenario. You could consider any
site-specific factors in developing a reasonable estimate of quantity released (e.g.,
the quantity that could be released from a sheared pipe in the time it would take to
shut off flow to the pipe).

Release Height. You may assume any appropriate release height for your
alternative scenarios. For example, you may analyze a scenario in which a regulated
substance would be released at a height well above ground level.

Wind Speed and Atmospheric Stability. You should use typical meteorological
conditions at your site to model alternative scenarios. To determine typical
conditions, you may need to obtain local meteorological data that are applicable to
your site. If you do not keep weather data for your site (most sources do not), you
may call another nearby source, such as an airport, or a compiler, such as the
National Weather Service, to determine wind speeds for your area. Your airport or
other source will be able to give you information on cloud cover.

ESTIMATING RELEASE RATES

Toxic Gases. To estimate a release rate for toxic gases, you may make any
appropriate assumptions based on conditions at your site and use any appropriate
model. EPA's RMP Offsite Consequence Analysis Guidance provides a simple
equation and chemical-specific data for estimating the release rate of a gas from a
hole in a vessel or pipe based on hole size, tank pressure, and chemical properties.
The size of the hole might be estimated from, for example, the hole size that would
result from shearing off a valve or pipe from a vessel.

Tank or pipe damage or failure resulting in the release of a gas liquefied under
pressure might be an appropriate alternative scenario at some sites. If such a release
would be possible at your site, you may need to consider a model or method that will
deal with this type of scenario.

You also should consider the duration of the release. EPA does not require you to
assume any specific time period for the release. You could estimate the release
Chapter 4
Offsite C onsequence A nalysis 4-16

duration based on the length of time it would take to stop the release, or you could
estimate a maximum duration based on a calculated release rate and the quantity in
the tank or pipes. If you estimate that a release of toxic gas would be stopped very
quickly, resulting in a "puff" rather than a plume, you may want to use a model that
deals with puff releases. EPA's RMP Offsite Consequence Analysis Guidance is not
appropriate for estimating distance to an endpoint for puff releases. EPA provides
information on various other models, including those suitable for analysis of puff
releases, at the Support Center for Regulatory Air Models:
[Link]

You may consider both passive and active mitigation in estimating release rates. For
gases, passive mitigation may include enclosed spaces. Active mitigation for gases
may include an assortment of techniques including automatic shutoff valves, rapid
transfer systems (emergency drainage), and water/chemical sprays. These mitigation
techniques have the effect of reducing either the release rate or the duration of the
release, or both. EPA's RMP Offsite Consequence Analysis Guidance includes
methods of accounting for mitigation. You also may use your knowledge or other
methods to account for mitigation.

Toxic liquids. For alternative releases of toxic liquids, you may consider any
scenario that would be reasonable for your site. For alternative release scenarios, you
are permitted to take credit for both passive and active mitigation systems, or a
combination if both are in place. For liquids, passive mitigation may include
techniques such as dikes and trenches. Active mitigation for liquids may include an
assortment of techniques including automatic shutoff valves, emergency drainage,
foam or tarp coverings, and water or chemical sprays. These mitigation techniques
have the effect of reducing either the quantity released into the pool or the
evaporation rate from the pool. EPA's RMP Offsite Consequence Analysis Guidance
discusses some methods of accounting for mitigation.

ESTIMATING DISTANCE TO THE ENDPOINT

For alternative releases, you may use any appropriate model (as discussed in 4.1) to
estimate the distance to the specified endpoint for an alternative release of a
regulated toxic substance. You may use site-specific conditions, including typical
weather conditions, and consider any site-specific factors appropriate to your
scenario. You must use the endpoints specified in part 68 Appendix A, as for the
worst-case analysis.

ALTERNATIVE RELEASES OF FLAMMABLE SUBSTANCES

Alternative release scenarios for flammable substances are somewhat more

complicated than for toxic substances because the consequences of a release and the
endpoint of concern may vary. For the worst case, the consequence of concern is a
vapor cloud explosion, with an overpressure endpoint. For alternative scenarios
involving fires rather than explosions, other endpoints than overpressure (e.g., heat
radiation) may need to be considered. The rule specifies endpoints for fires based on
the heat radiation level that may cause second degree burns from a 40-second
exposure and the lower flammability limit (LFL), which is the lowest concentration
 Chapter 4
4-17 Offsite C onsequence A nalysis

in air at which a substance will burn. Some possible scenarios involving flammable
substances are discussed below.

� Vapor cloud fires (flash fires) may result from dispersion of a cloud of
flammable vapor and ignition of the cloud following dispersion. Such a fire
could flash back and could represent a severe heat radiation hazard to
anyone in the area of the cloud. Vapor cloud fires may be modeled using air
dispersion modeling techniques to estimate distances to a concentration
equal to the LFL.

� A pool fire, with potential radiant heat effects, may result from a spill of a
flammable liquid. The endpoint for this type of fire, as listed in the rule, is a
radiant heat level of 5 kilowatts per square meter (kW/m2) for 40 seconds; a
40-second exposure to this heat level could cause second degree burns.

� A boiling liquid, expanding vapor explosion (BLEVE), leading to a

fireball that may produce intense heat, may occur if a vessel containing
flammable material ruptures explosively as a result of exposure to fire. Heat
radiation from the fireball is the primary hazard; vessel fragments and
overpressure from the explosion also can result. BLEVEs are generally
considered unlikely events; however, if you think a BLEVE is possible at
your site, you should estimate the distance at which radiant heat effects
might lead to second degree burns, since that is the effect of concern
underlying the rule’s endpoint for fires. The point of offsite consequence
analyses is to determine how far away from the point of release effects of
concern could occur, so you should estimate the distance for BLEVEs even
though one would not last for 40 seconds. For BLEVEs, you would need to
estimate the radiant heat level at which exposure for the duration of the
BLEVE would cause second degree burns. You then would need to estimate
the distance to that heat level. You also may want to consider models or
calculation methods to estimate effects of vessel fragmentation, although you
are not required to analyze such effects.

� For a vapor cloud explosion to occur, rapid release of a large quantity of

flammable material, turbulent conditions (caused by a turbulent release or
congested conditions in the area of the release, or both), and other factors are
generally necessary. Vapor cloud explosions generally are considered
unlikely events; however, if conditions at your site are conducive to vapor
cloud explosions, you may want to consider a vapor cloud explosion as an
alternative scenario. The 1 psi overpressure endpoint still applies to a vapor
cloud explosion for purposes of analyzing an alternative scenario, but you
could use less conservative assumptions than for the worst-case analysis,
including any reasonable estimate of the quantity in the cloud and the yield
factor. A vapor cloud deflagration, involving lower flame speeds than a
detonation and resulting in less damaging blast effects, is more likely than a
detonation. You may assume a vapor cloud deflagration for the alternative
scenario, if you think it is appropriate, and use the radiant heat endpoint
(adjusted for duration).
Chapter 4
Offsite C onsequence A nalysis 4-18

� A jet fire may result from the puncture or rupture of a tank or pipeline
containing a compressed or liquefied gas under pressure. The gas can form
a jet that discharges into the air in the direction of the hole; the jet then may
ignite. Jet fires could contribute to BLEVEs and fireballs if they impinge on
tanks of flammable substances. A large horizontal jet fire may have the
potential to pose an offsite hazard. You may want to consider a jet fire as an
alternative scenario, if appropriate for your site.

MODELING ASSUMPTIONS

Quantity. EPA has not specified any assumptions you must make concerning
quantity released for alternative scenario analysis for flammable substances. You
may consider any site-specific factors in developing a reasonable estimate of
quantity released, as for toxic substances (e.g., the quantity that could be released
from a ruptured pipe in the time it would take to shut off flow to the pipe).

Release Height. You may assume any appropriate release height for your
alternative scenarios for flammable substances.

Wind Speed and Atmospheric Stability. Meteorological conditions may have little
effect on some scenarios for flammable substances (e.g., vapor cloud explosions and
BLEVEs), but may have a relatively large effect on others (e.g., a vapor cloud fire
resulting from downwind dispersion of a vapor cloud and subsequent ignition). You
should use typical meteorological conditions at your site to model appropriate
alternative scenarios. To determine typical conditions, you may need to obtain local
meteorological data that are applicable to your site, as discussed above.

ESTIMATING RELEASE RATES

Flammable Gases. To estimate a release rate for flammable gases, you may make
any appropriate assumptions based on conditions at your site. You may consider the
effects of both passive and active mitigation systems. The methods provided in
EPA's RMP Offsite Consequence Analysis Guidance for rate of release of a gas from
a hole in a vessel or pipe for toxic gases also can be used for flammable gases.
Chemical-specific data are provided for flammable gases, to be used along with hole
size and tank pressure to estimate release rates.

Flammable liquids. For alternative releases of flammable liquids, you may

consider any scenario that would be reasonable for your site. You are permitted to
take credit for both passive and active mitigation systems, or a combination if both
are in place, as for toxic liquids. You could consider release of the liquid into a pool
and release to air by pool evaporation, if you consider this to be a reasonable
scenario.

If evaporation of a flammable liquid from a pool is an appropriate assumption for

your alternative scenario, you can use any scientifically appropriate method to
estimate the evaporation rate.
 Chapter 4
4-19 Offsite C onsequence A nalysis

ESTIMATING DISTANCE TO THE ENDPOINT

You may use any appropriate model to estimate the distance to the specified
endpoint for alternative scenarios for regulated flammable substances. Several
possible consequences of releases of flammable substances are discussed below.

Vapor cloud fire. You may use any appropriate model to estimate distances for a
vapor cloud fire. The LFL endpoint, specified in the rule, would be appropriate for
vapor cloud fires. You may use air dispersion modeling to estimate the maximum
distance to the LFL. You may want to consider, however, whether it is likely that a
flammable gas or vapor could disperse to the maximum distance to the LFL before
reaching an ignition source. The actual dispersion distance before ignition might be
much shorter than the maximum possible distance.

Pool fire. Any appropriate model may be used for pool fires of flammable liquids.
The applicable endpoint specified in the rule is the heat radiation level of 5 kW/m2 .

BLEVE. If a fireball from a BLEVE is a potential release scenario at your site, you
may use any model or calculation method to estimate the distance to a radiant heat
level that can cause second degree burns (a heat "dose" equivalent to the specified
radiant heat endpoint of 5 kW/m2 for 40 seconds).

Vapor cloud explosion. If you have the potential at your site for the rapid release of
a large quantity of a flammable vapor, particularly into a congested area, a vapor
cloud explosion may be an appropriate alternative release scenario. For the
alternative analysis, you may estimate any reasonable quantity of flammable
substance in the vapor cloud. The endpoint for vapor cloud explosions is 1 psi, as
for the worst case; however, a smaller yield factor may be used for the alternative
scenario analysis.

NUMBE R OF SCENARIOS

You are required to analyze at least one alternative release scenario for each listed
toxic substance you have in a Program 2 or Program 3 process above its threshold
quantity. If you have the same substance above the threshold in several processes or
locations, you need only analyze one alternative scenario for it. You also are
required to analyze one alternative release scenario representing all regulated
flammable substances in Program 2 or 3 processes; you do not need to analyze an
alternative scenario for each flammable substance above the threshold. For example,
if you have five listed substances — chlorine, ammonia, hydrogen chloride,
propylene, and acetylene — above the threshold in Program 2 or 3 processes, you
will need to analyze one alternative scenario each for chlorine, ammonia, and
hydrogen chloride (toxics) and a single alternative scenario to cover propylene and
acetylene (flammable substances).

No alternative scenario analysis is required for regulated substances in Program 1

processes. If worst-case scenario analysis for a process shows no public receptors
within the distance(s) to the applicable endpoint(s), and the process meets the other
Program 1 criteria, you do not have to carry out an alternative scenario analysis for
the process.
Chapter 4
Offsite C onsequence A nalysis 4-20

In addition, no alternative scenario analysis is required for any process that does not
contain more than a threshold quantity of a regulated substance, even if you believe
such a process is a likely source of a release.

4.4 ESTIMATING OFFSITE RECEPTORS

The rule requires that you include in your RMP an estimate of the residential
populations within the geographical areas that could be affected by the hypothetical
worst-case or alternative releases you have analyzed. The geographical area is the
area within the circle defined by the endpoint for your worst-case and alternative
release scenarios (i.e., the center of the circle is the point of release and the radius is
the distance to the endpoint). In addition, you must report in the RMP whether
certain types of public receptors and environmental receptors are within the circles.

RESIDENTIAL POPULATIONS

To estimate residential populations, you may use the most recent Census data or any
other source of data that you believe is more accurate. You are not required to
update Census data or conduct any surveys to develop your estimates. Census data
are available in public libraries and in the LandView system, which is available on
DVD (see box below). The rule requires that you estimate populations to two
significant digits. For example, if there are 1,260 people within the circle, you may
report 1,300 people. If the number of people is between 10 and 100, estimate to the
nearest 10. If the number of people is less than 10, provide the actual number.

Census data are presented by Census tract. If your circle covers only a portion of the
tract, you should develop an estimate for that portion. The easiest way to do this is
to determine the population density per square mile (total population of the Census
tract divided by the number of square miles in the tract) and apply that density figure
to the number of square miles within your circle. Because there is likely to be
considerable variation in actual densities within a Census tract, this number will be
approximate. The rule, however, does not require you to correct the number.

OTHER PUBLIC RECEPTORS

Other public receptors must be noted in the RMP (see the discussion of public
receptors in Chapter 2). If there are any schools, residences, hospitals, prisons,
public recreational areas or arenas, or commercial or industrial areas within the
circle, you must report that. You are not required to develop a list of all public
receptors; you must simply check off that one or more such areas is within the circle.
Most receptors can be identified from local street maps.

ENVIRONM ENTAL RECEPTORS

Environmental receptors are defined as natural areas such as national or state parks,
forests, or monuments; officially designated wildlife sanctuaries, preserves, refuges,
or areas; and Federal wilderness areas. Only environmental receptors that can be
identified on local U.S. Geological Survey (USGS) maps (see box below) need to be
considered. You are not required to locate each of these specifically. You are only
required to check off in the RMP which specific types of areas are within the circle.
 Chapter 4
4-21 Offsite C onsequence A nalysis

If any part of one of these receptors is within your circle, you must note that in the
RMP.

Important: The rule does not require you to assess the likelihood, type, or severity
of potential impacts on either public or environmental receptors. Identifying them as
within the circle simply indicates that they could be adversely affected by the
release.

HOW TO OBTAIN CENSUS DATA AND LAND VIEW ®

Census data can be found in publications of the Bureau of the Census, available in public libraries,
including County and City Data Book.

LandView ® is a desktop mapping system that includes database extracts from EPA, the Bureau of
the Census, the U.S. Geological Survey. These databases are presented in a geographic context on
maps that show jurisdictional boundaries, detailed networks of roads, rivers, and railroads, census
block group and tract polygons, schools, hospitals, churches, cemeteries, airports, dams, and other
landmark features. LandView lets you determine the population within any radius, anywhere in the
U.S.

For PC and Mac. $129 for 2 DVD set, covers entire country

U.S. Department of Commerce

Bureau of the Census
P.O. Box 277943
Atlanta, GA 30384-7943
Phone: (301) 763-INFO (4636) (Customer Services -- orders)
Phone: (301) 763-1128 (Geography Staff -- content)
[Link]

Further information on Census data is available at the Bureau of the Census web site at
[Link].
Chapter 4
Offsite C onsequence A nalysis 4-22

HOW TO OBTAIN USGS MAPS

The production of digital cartographic data and graphic maps comprises the largest component of the
USGS National Mapping Program. The USGS's most familiar product is the 1:24,000-scale
Topographic Quadrangle Map. This is the primary scale of data produced, and depicts greater detail
for a smaller area than intermediate-scale (1:50,000 and 1:100,000) and small-scale (1:250,000,
1:2,000,000 or smaller) products, which show selectively less detail for larger areas.

To order maps, publications, and other miscellaneous products on paper and compact disc media
from the USGS by mail, send your order form and payment to:

USGS Information Services

Box 25286
Denver, CO 80225

By fax, call 303-202-4693 and transmit the order form.

To order digital maps, digital cartographic and geographic data, aerial photographs, and satellite
imagery products from the USGS by mail, send your order form and payment to:
USGS EROS Data Center
47914 252nd Street
Sioux Falls, SD 57198

By fax, call 605-594-6589 and transmit the order form.

Online, submit your order through the USGS on-line ordering site, The USGS Store at:
[Link]

For more information, visit the USGS Earth Science Information Center on the Internet at
[Link] or for ordering assistance, call 1-888-ASK-USGS. For additional information,
contact any USGS Earth Science Information Center or call 1-800-USA-MAPS.
 Chapter 4
4-23 Offsite C onsequence A nalysis

Qs and As
OFFSITE CONSEQUENCE ANALYSIS

Q. How close must a stationary source be to a weather station for that station's data to be applicable
to the stationary source?

A. EPA has not set specific distance limits, but will allow owners and operators to use reasonable
judgement in determining whether data from a weather station is applicable to the stationary source.
Factors such as topography and distance between the stationary source and a weather station should
be taken into consideration when evaluating the applicability of the weather station's data to the
stationary source.

Q. Must air dispersion models that are used to analyze worst-case release scenarios be able to
account for multiple vessels and how those vessels could impact one another in the event of an
accidental release?

A. No. Models used for worst-case release scenario analysis do not need to consider compounding
effects of accidental releases from multiple vessels because worst-case release is defined by the rule
as a single vessel or process line failure that will result in the greatest distance to an endpoint.

Q. If the estimated population changes, would the RMP have to be updated?

A. No. Changes in U.S. Census data do not necessitate a correction or update of the RMP.
However, all updates to the RMP should use the most recent U.S. Census data.

Q. What if a flammable event has a different time duration than the 5 kw/m2 for 40 seconds?

A. EPA recognizes that flammable events may occur for a different amount of exposure time.
Therefore, the owner or operator should determine the distance to an equivalent exposure - e.g. if the
flammable event occurs for 20 seconds, determine the distance to an equivalent exposure level.

Q. Could positive buoyancy models be used?

A. Yes, provided there is a basis for use and the owner or operator explains the rationale for use of
positive buoyancy models.
Chapter 4
Offsite C onsequence A nalysis 4-24

4.5 Examples of Worst Case Scenarios

EXAMPLE < SOURCE A

Source A, a retail operation that supplies ammonia, has one covered process: a 200-ton ammonia
storage tank. Source A carries out worst-case consequence analyses for this process, with the
following results: For 400,000 pounds of anhydrous ammonia, the distance to the specified endpoint
(0.14 mg/L) is estimated as more than 10 miles.

Residences and a business center are located within 0.15 miles of the facility; therefore the regulated
process is not eligible for Program 1. Source A must report the results of the worst-case analysis in
the RMP.

EXAMPLE < SOURCE B

Source B is a medium-sized metal products manufacturer with two processes containing regulated
toxic substances above their thresholds: a tank storing 50,000 pounds of 37 percent hydrochloric acid
for use in plating processes and five interconnected, one-ton tanks of chlorine used in a wastewater
treatment plant. Only one worst-case analysis is required for toxic substances for Program 2 and
Program 3 processes; because of the greater toxicity and volatility of chlorine, Source B expects that
a worst-case release of chlorine would result in the greatest distance to the endpoint. Source B does
not believe the hydrochloric acid process would be eligible for Program 1 because of the proximity
of public receptors (including workers at an adjacent industrial facility), and, therefore, only carries
out the worst-case analysis for the chlorine process. A distance of 2.80 miles to the endpoint is
estimated for a release of 2,000 pounds of chlorine. Since there are public receptors within that
distance, Source B must report this worst-case analysis in the RMP.

EXAMPLE < SOURCE C

Source C is an inorganic chemical manufacturer with two covered processes: a tank containing 10
tons of 70 percent hydrofluoric acid solution and ten one-ton tanks of chlorine on a rack for
wastewater treatment. Source C must carry out one worst-case analysis for regulated toxic
substances for Program 2 and Program 3 processes. Because the toxic endpoint of chlorine is lower
than that of hydrofluoric acid, and because the release rate will probably be greater for a gas than a
solution, Source C decides to carry out the analysis for chlorine as the required worst-case analysis
for toxic substances. Source C believes the hydrofluoric acid process may be eligible for Program 1
and, therefore, decides to do a worst-case analysis for this process as well. Results of the worst-case
analyses for these two processes are:

� 2.80 miles for 2,000 pounds of chlorine

� 1 mile for 20,000 pounds of 70 percent hydrofluoric acid (released in a diked area)

Homes and businesses are located less than a mile from either process; therefore, neither process is
eligible for Program 1. Source C must report the results of the worst-case analyses for chlorine.
 Chapter 4
4-25 Offsite C onsequence A nalysis

EXAMPLE < SOURCE D

Source D is a large chemical manufacturer with 11 regulated substances above their threshold
quantities, including three flammable substances and eight toxic substances. The processes
containing flammable substances are: three 18,000-gallon tanks containing 26,000 pounds of
ethylene, 66,000 pounds of propylene, and 65,000 pounds of propane; the propane, however, is used
as a fuel and is not subject to the rule. The largest quantities of toxic substances in processes are:
25,000 pounds of toluene diisocyanate (TDI), 100,000 pounds of chloroform, 25,000 pounds of
anhydrous hydrogen chloride, 20,000 pounds of chlorine, 80,000 pounds of epichlorohydrin, 100,000
pounds of methyl chloride, 10,000 pounds of hydrogen cyanide, and 1,000 pounds of phosgene. For
the RMP, Source D has to report one worst-case analysis for flammable substances and one for toxic
substances; however, Source D believes the processes containing flammable substances may be
eligible for Program 1 and, therefore, chooses to carry out a worst-case analysis for each of these
processes. In addition, Source D believes the processes containing TDI and chloroform may be
eligible for Program 1, because of the low volatility of TDI and the relatively low toxicity of
chloroform, and decides to carry out analyses to determine eligibility. Source D is not sure which of
the other processes containing toxic substances will give the greatest distance to the endpoint;
therefore, it conducts screening analyses for all these processes. The worst-case distances for vapor
cloud explosions of the flammable substances are:
� 0.24 miles for 26,000 pounds of ethylene;
� 0.32 miles for 66,000 pounds of propylene.

The worst-case distances to the endpoints for the toxic substances are:
� 0.06 miles for 25,000 pounds of TDI;
� 0.49 miles for 100,000 pounds of chloroform;
� 4.8 miles for 25,000 pounds of hydrogen chloride;
� 10 miles for 20,000 pounds of chlorine;
� 2.2 miles for 80,000 pounds of epichlorohydrin;
� 2.0 miles for 100,000 pounds of methyl chloride;
� 5.2 miles for 10,000 pounds of hydrogen cyanide; and
� 11 miles for 1,000 pounds of phosgene.

The processes containing ethylene and propylene are located 500 yards (0.28 miles) from a river (0.5
miles wide). The distance to the endpoint for these two processes does not extend beyond the river,
which is not a recreational area; the processes are eligible for Program 1 (having met the other
criteria). The distances to the endpoint for the TDI process do not reach public receptors in any
direction; therefore, this process is also eligible for Program 1.

Source D reports the worst-case analysis results for ethylene, propylene, and TDI to demonstrate
eligibility for Program 1. Since propylene was the “worst” worst-case scenario for all regulated
flammables at the facility, the source does not have to report any other worst-case scenario results
for flammable substances. The source does report the results for phosgene as the required worst-case
analysis for toxic substances.
Chapter 4
Offsite C onsequence A nalysis 4-26

4.6 Examples of Alternative Releases

EXAMPLE < SOURCE A

Source A, a retail operation that supplies ammonia and propane, has one covered process: an
ammonia storage tank containing 400,000 pounds of anhydrous ammonia (a regulated toxic
substance). The worst-case consequence analyses for this process indicated it is not eligible for
Program 1. Source A must carry out and report an alternative scenario analysis for the process. Any
reasonable and defensible scenarios can be analyzed for the process, provided they are more likely
than the worst-case scenario and reach an off-site endpoint (unless no such scenario exists). The
source must be able to explain its choice of scenarios.

EXAMPLE < SOURCE B

Source B is a medium-sized metal products manufacturer with two covered processes containing
regulated toxic substances: a chlorine wastewater treatment plant with 10,000 pounds of chlorine and
a tank containing 50,000 pounds of 37 percent hydrochloric acid. Because of the proximity of public
receptors, neither of these processes is eligible for Program 1. Source B must carry out and report an
alternative scenario analysis for each of the two regulated substances. Source B may analyze any
scenarios that are reasonable for the site and processes, provided they are more likely than the worst-
case scenario and reach an off-site endpoint (unless no such scenario exists). The source must be
able to explain its choice of scenarios.

EXAMPLE < SOURCE C

Source C is an inorganic chemical manufacturer with two covered processes, one containing 20,000
pounds of chlorine and the other containing 20,000 pounds of 70 percent hydrofluoric acid. Source
C’s worst-case analyses indicated that these processes are not eligible for Program 1. Source C must
carry out and report an alternative scenario analysis for each of these substances. The scenarios must
be more likely than the worst-case scenario and reach an off-site endpoint (unless no such scenario
exists). Within these constraints, the scenarios may be developed based on any reasonable and
defensible assumptions. The source must be able to explain its choice of scenarios.
 Chapter 4
4-27 Offsite C onsequence A nalysis

EXAMPLE < SOURCE D

Source D is a large chemical manufacturer with covered processes containing two regulated
flammable substances and eight regulated toxic substances. The worst-case analyses showed that the
processes containing the flammable substances (ethylene and propylene) are eligible for Program 1.
For flammable substances, Source D need not carry out and report an alternative scenario analysis.

The worst-case analyses showed that the process containing 25,000 pounds of the toxic substance
toluene diisocyanate (TDI) also is eligible for Program 1; therefore, Source D does not need to carry
out an alternative scenario analysis for TDI. Source D must carry out and report an alternative
scenario analysis for each regulated toxic substance in a covered non-Program 1 process; thus,
scenarios must be developed and analyzed for hydrogen chloride, chlorine, epichlorohydrin, methyl
chloride, hydrogen cyanide, chloroform, and phosgene. If the substances are found in more than one
vessel, the analysis should be conducted with respect to the vessel that presents the greatest relative
risk of a release. Analyses of each vessel are not needed. Source D can develop any reasonable
scenarios for these substances that are more likely than the worst-case scenario and reach an off-site
endpoint (unless no such scenario exists), and must be able to explain its choice of scenario.