Dispersion Modelling for Toxic Gas Release Step-by-Step

If you are moving into process safety or QRA work, dispersion modelling is one of the first technical skills you will need to get comfortable with. It is not just a software exercise. It tells you where a toxic gas cloud will travel after a release, how concentrated it will be at different distances, and whether people in the surrounding area are at risk.

Most fresh graduates stepping into safety engineering roles underestimate how much judgement goes into setting up a dispersion model correctly. The software output is only as reliable as the inputs you define. Get the source term wrong, pick the wrong stability class, or misread the consequence distance, and your entire risk assessment is built on shaky ground.

This guide walks through the full process, from understanding what a release scenario looks like to interpreting real model outputs. Whether you are preparing for a QRA role or trying to understand what your safety team produces, this is where to start.

What Is Dispersion Modelling and Why It Matters in Process Safety

Dispersion modelling is a mathematical method used to predict how a hazardous gas or vapour spreads through the atmosphere after an accidental release. It calculates the concentration of the toxic substance at various distances and directions from the release point, under defined weather conditions.

In process safety work, dispersion modelling sits inside a broader consequence analysis workflow. After identifying a loss of containment scenario, say, a pipeline rupture or a valve failure, the next question is, ‘What happens to the released material?’ Does it ignite? Does it drift toward a populated area? How far does the toxic cloud travel before concentrations drop to safe levels?

These are not theoretical questions. Regulatory bodies and site operators use dispersion modelling results to set land use planning zones, design emergency response plans, justify safe distances for control rooms and muster points, and satisfy requirements under process safety management frameworks.

For engineers working under process safety management frameworks, OSHA’s PSM standard provides the regulatory foundation that makes consequence analysis and dispersion modelling a compliance requirement on major hazard sites.

How a Toxic Gas Release Actually Happens: Source Term Basics

Before running any dispersion model, you need to define the source term. This is the starting point, and it is where many junior engineers make their first mistake by jumping straight into the software without properly characterising the release.

The source term describes what is being released, at what rate, and in what form. Dispersion modelling software like PHAST takes these inputs and uses them to calculate how the cloud behaves as it moves downwind.

Jet Release vs Pool Evaporation

A jet release happens when pressurised fluid escapes through a hole or rupture. Think of a gas pipeline failure or a pressurised vessel leak. The material exits at high velocity, and the initial momentum drives the dispersion behaviour before atmospheric dilution takes over.

A pool evaporation scenario occurs when a liquid releases onto the ground and then evaporates. Liquefied ammonia or chlorine spilt onto the ground does not instantly become a gas cloud; it evaporates progressively, and the evaporation rate depends on surface area, ambient temperature, and solar radiation.

Getting this distinction right matters because the two scenarios produce very different concentration profiles downwind.

Key Parameters That Define a Source Term

• Release hole size and geometry
• Operating pressure and temperature at the release point
• Physical state of the material (gas, liquid, or two-phase)
• Duration of release (instantaneous or continuous)
• Molecular weight and vapour pressure of the substance
• Height of release above grade

Skipping any of these or using conservative estimates without justification can produce results that are either unrealistically severe or dangerously optimistic.

Understanding Atmospheric Conditions and Stability Classes

One thing that surprises many engineers new to dispersion modelling is how significantly the weather affects the results. The same release scenario can produce a toxic plume that travels 200 metres under one set of conditions and over 2 kilometres under another.

Pasquill-Gifford stability classes are the standard framework used to characterise atmospheric turbulence in dispersion modelling. They range from Class A (very unstable, strong daytime convection) through to Class F (very stable, calm night-time conditions with low wind).

Class A and B conditions occur during sunny daytime periods with light winds. Atmospheric mixing is strong, which dilutes the gas cloud quickly. Class F conditions typically occur on clear nights with very light winds. The atmosphere is stable, mixing is minimal, and a toxic cloud can travel much further at harmful concentrations.

Wind speed is equally important. Higher wind speeds generally dilute the cloud faster in the crosswind direction but transport it further downwind. Lower wind speeds produce a narrower, more concentrated plume that moves slowly.

In practice, process safety studies typically run dispersion modelling across multiple stability class and wind speed combinations, often F/1.5 ms⁻¹ for worst-case toxic scenarios and D/5 ms⁻¹ for more typical conditions. Regulators sometimes specify which combinations to use for land use planning assessments.

Ignoring the sensitivity of results to weather inputs is one of the most common oversights seen in early-career engineers handling dispersion modelling for the first time.

Step-by-Step Dispersion Modelling Workflow

The actual modelling process follows a structured sequence. Rushing through it or skipping steps leads to results that cannot be defended in a review or regulatory submission.

Step 1: Define the Release Scenario

Start by identifying the specific loss of containment scenario you are modelling. Is it a full-bore rupture, a small leak from a flange, or a catastrophic vessel failure? The scenario selection usually comes from a HAZOP study, a failure frequency database, or an engineering judgement based on the system design.

Be specific. Define the release location, the equipment involved, and the isolation assumptions. Can the release be isolated remotely? Does it continue for 10 minutes or until the vessel empties?

Step 2: Establish Source Term Inputs

Using the scenario defined in Step 1, calculate or estimate the release rate. For pressurised gas systems, this typically involves orifice flow equations. For liquids, you need to account for flash fraction if the substance is above its normal boiling point.

PHAST has built-in source term modules that can calculate release rates directly from process conditions if you input the fluid properties, hole size, and upstream pressure. Most engineers use these built-in calculators, but knowing what is happening in the background helps you spot errors.

Step 3: Select Atmospheric Conditions

Based on the study requirements, select the stability class and wind speed combinations to model. For worst-case toxic consequence distances, F/1.5 ms⁻¹ is most commonly used. For risk-based studies, you may need to model multiple weather conditions weighted by their frequency of occurrence.

Step 4: Run the Dispersion Model

Input all parameters into the software and run the model. PHAST generates concentration profiles as a function of distance, along with toxic dose calculations based on the selected toxicological endpoints.

Check the output for obvious anomalies before proceeding. Unusually short or long consequence distances should always prompt a review of inputs.

Step 5: Interpret Toxic Dose and Consequence Distances

The output is not just a number. Understand what toxic endpoint you are assessing against. Are you calculating distances to IDLH (Immediately Dangerous to Life or Health), ERPG-2, or a probit-based 1% lethality level? Each serves a different purpose in the overall risk assessment.

Dispersion Modelling Software Used in Industry

Several software tools are used across industry for dispersion modelling. The choice of tool often depends on the study type, company standards, and regulatory requirements in the operating region.

Software	Developer	Primary Use	Industry Adoption
PHAST	DNV	Full consequence modelling, QRA	Oil and gas, petrochemical, chemical
SAFETI	DNV	QRA with risk contour mapping	Major hazard facility assessments
ALOHA	NOAA / EPA	Emergency response planning	Public sector, emergency services
CAMEO	EPA	Chemical emergency planning	Regulatory and response teams

PHAST is the most widely used tool in professional process safety and QRA work. It handles jet releases, pool fires, vapour cloud explosions, and toxic dispersion within a single platform. If you are targeting a career in consequence analysis or QRA, PHAST proficiency is what most employers look for.

SAFETI extends PHAST by adding risk integration across multiple scenarios and generating individual and societal risk outputs. These are the risk contours you see on land use planning maps around major hazard sites.

ALOHA and CAMEO are more commonly used by emergency response teams and regulatory bodies rather than design engineers. They are simpler to use but lack the depth needed for formal QRA submissions.

Reading and Interpreting Dispersion Modelling Results

Getting the model to run is one thing. Reading the output correctly is where judgement comes in and where many engineers struggle early in their careers.

Dispersion modelling results typically show concentration versus distance plots, toxic dose contours, and consequence distances to specified toxic endpoints. The three endpoints most commonly used in process safety are:

• ERPG-2: the maximum airborne concentration below which nearly all individuals could be exposed for up to one hour without experiencing irreversible health effects
• ERPG-3: the concentration threshold associated with life-threatening effects
• IDLH: immediately dangerous to life or health, commonly used in emergency response planning

The consequence distance is the downwind distance at which concentrations drop below the selected endpoint. This distance feeds directly into risk contour calculations and emergency planning zones.

One mistake engineers commonly make is treating the consequence distance as a hard boundary. In reality, the model produces a probability distribution of outcomes. Wind direction variability, terrain effects, and building wake effects all influence where the cloud actually goes. Real plant sites are rarely as open and flat as the model assumes.

During a QRA project for a refinery expansion, it is not unusual to find that consequence distances for an ammonia release from a refrigeration system extend significantly beyond the site boundary under F/1.5 ms⁻¹ conditions. That finding has direct implications for land use planning, evacuation zones, and design decisions on system isolation.

Always document your input assumptions. If a result is challenged in a regulatory review, the first thing the reviewer asks for is the basis of your source term and weather selection.

How Dispersion Modelling Feeds into QRA Risk Contours

Dispersion modelling does not stand alone. In a full QRA study, the consequence distances it produces are only one half of the risk picture. The other half is frequency – how often a given release scenario is likely to occur. Combine consequence with frequency, and you get risk.

Once dispersion modelling has established the toxic consequence distance for a specific scenario, that distance gets translated into a fatality probability at each location around the facility. This is done using a probit function or a simplified harm model that converts concentration and exposure time into a percentage probability of fatality.

That probability, combined with the release frequency from a failure frequency database, gives the individual risk at that point. Repeat this across all modelled scenarios – different hole sizes, different substances, and different weather conditions – and you build up a complete individual risk contour map around the site.

The methodology for integrating consequence and frequency into individual risk contours is detailed in the CCPS Guidelines for Chemical Process Quantitative Risk Analysis, which remains the primary reference document used by QRA practitioners globally.

Career Scope for Engineers Skilled in Dispersion Modelling

Process safety engineering is one of the more specialised and consistently in-demand disciplines across oil and gas, petrochemical, LNG, and chemical manufacturing. Dispersion modelling sits at the heart of several key roles within this space.

Engineers with hands-on dispersion modelling experience typically work in roles such as:

• Process Safety Engineer consequence analysis, HAZOP support, safety case development
• QRA Engineer full risk quantification, including frequency analysis and consequence modelling
• HSE Consultant: third-party safety studies, regulatory submissions, emergency response planning
• Loss Prevention Engineer facility risk assessments, occupied building studies, layout reviews

The skill is not limited to consultancy. Operating companies, EPC contractors, and major hazard operators all need engineers who can run and interpret dispersion modelling results internally.

What makes this skill particularly valuable is that it connects directly to decision-making. A consequence distance from a dispersion model determines where a control room can be sited; whether a residential area falls within a planning consultation zone; and what emergency response resources need to be in place.

For engineers looking to build structured competency in consequence modelling and QRA workflows, the Advanced Quantitative Risk Assessment (QRA) Masterclass with PHAST & Safeti provides hands-on training with the industry-standard tools used in live project environments. Engineers wanting broader grounding in process safety methodology can also explore the Comprehensive Training in Process & Technical Safety Study Online Training, which covers the full safety study lifecycle from HAZOP through to risk assessment.

Conclusion

Dispersion modelling is not a black-box exercise you hand off to software and accept the output uncritically. Every result depends on the quality of the inputs, the release scenario, the source term, the atmospheric conditions and the engineering judgement applied at each step.

For engineers entering process safety or QRA roles, this is a skill worth developing with genuine depth. Understanding what the numbers mean, where they can mislead you, and how regulatory bodies interpret them is what separates a competent modeller from someone who can simply press run.

Start with the fundamentals covered here. Build your understanding of source terms and atmospheric behaviour before worrying about software shortcuts. Once the underlying logic is clear, tools like PHAST become straightforward to use and far more useful in real project environments.

FAQs