Dispersion Modelling Methodology sits at the intersection of process safety, environmental protection, and emergency planning, giving operators a structured way to predict how gases or aerosols move and dilute after an accidental or planned release. In the UAE’s high-hazard sectors — from refineries and gas processing plants to chemical terminals and power generation — understanding how a cloud of H2S, ammonia, chlorine, LNG vapor, or hydrocarbon mixture evolves in real terrain and weather is central to protecting people, safeguarding assets, and demonstrating regulatory compliance. This article explains the technical building blocks, the practical realities in facilities, and the decisions that dispersion results are designed to inform.
Who Does This Affect?

The knowledge and decisions tied to atmospheric dispersion extend across multiple teams. While process safety engineers often lead consequence modelling as part of Quantitative Risk Assessment (QRA) or Bowtie analysis, many stakeholders depend on the outcomes:
- HSE and process safety leadership using contours and concentration-time predictions to frame ALARP judgements and set design targets.
- Operations and maintenance teams who need to understand where gas may accumulate during maintenance vents, purge activities, or abnormal conditions.
- Emergency response planners shaping shelter-in-place strategies, muster points, and evacuation routes, including interactions with Civil Defence requirements.
- Engineering disciplines specifying ventilation rates, HVAC pressurization, and control room integrity consistent with API RP 752/753 and IEC 61511 lifecycle considerations.
- Project and EPC teams integrating dispersion outcomes in layout, separation distances, and flare design — especially for dense gases and low-temperature releases.
- Compliance officers preparing submissions for ADNOC group standards, OSHAD SF, Dubai Municipality Environmental permits, and other authority-having-jurisdiction approvals.
Because the modelling informs both design and operations, it becomes a shared reference for technical assurance, risk communication, and incident preparedness. The same set of results may guide a SIL/LOPA discussion on gas detection voting thresholds while also underpinning a community impact assessment for a storage terminal expansion.
Why This Topic Matters

Recent safety expectations in the region have converged around a few core goals: reduce the probability and severity of acute releases, improve control and mitigation layers, and plan credible emergency responses backed by data. Dispersion is the thread that connects source-term assumptions to the consequences experienced by people and assets. Without a defensible approach, organizations risk mis-sizing gas detection arrays, misplacing muster points, underestimating building pressurization needs, or overstating offsite risk in ways that can stall permitting and undermine trust.
Dispersion predictions inform multiple formal processes. In QRAs aligned with ADNOC HSE requirements, consequence modelling establishes individual risk contours and potential loss-of-life estimates. In hazard studies such as HAZID, HAZOP, and Bowtie reviews, they help visualize barriers and escalation pathways. For regulatory compliance, they underpin submissions to Civil Defence and environmental authorities, supporting arguments on hazardous area classification interfaces, ERT readiness, and land-use planning buffers.
Beyond governance, the topic matters because physics is unforgiving. Dense, cold hydrocarbon releases may slump and travel far under stable night-time conditions, even when intuition says a cloud “should” dissipate. Building wakes can recirculate a toxic plume into an upwind air intake. Sea breeze circulations common to coastal UAE sites can rotate hazard footprints over the course of a day. Dispersion modelling makes these less obvious realities explicit, so teams can put controls in the right place the first time.
Core Elements of Dispersion Modelling Methodology

To be decision-grade, a dispersion study blends physical realism with practical simplicity. The following elements define a technically sound approach, regardless of the software chosen.
1) Source term definition

All models begin with a source: what is released, at what rate, for how long, at what temperature, and through what geometry. The source term links directly to process design, equipment sizing, and safeguards:
- Fluid properties: composition, density, molecular weight, heat capacity, vapor pressure, condensation and solidification potential.
- Release characteristics: hole size, upstream pressure/temperature, two-phase flashing, choked vs subsonic flow, relief and blowdown dynamics, jet momentum and buoyancy.
- Time dependence: instantaneous (puff), continuous (plume), or time-varying releases tied to blowdown curves or control actions.
Getting the source term wrong propagates error into every subsequent prediction. Close alignment with HAZOP deviations, PSV/BV datasheets, and control logic (ESD, blowdown timings) is essential, especially where IEC 61511 safety functions may truncate release duration.
2) Meteorology and site representativeness

Wind speed, direction, stability class, turbulence, and temperature stratification govern how quickly a cloud dilutes and where it travels. In the UAE, several patterns matter:
- Sea breeze and land breeze cycles around the Gulf coast that can steer plumes on diurnal timescales.
- Stable night-time conditions with light winds that support longer travel distances for dense gases.
- Thermal convection during hot afternoons that increases vertical mixing.
- Dust-laden or humid air influencing radiative cooling and, in some cases, buoyancy behavior for hot jets.
Ideally, an onsite met mast provides a year or more of data to build wind roses, stability distributions, and roughness length estimates. When not available, careful proxy selection and transparent uncertainty discussion are critical. Averaging times must be appropriate for endpoints: for example, 1-minute or 10-minute toxic averages versus instantaneous LFL envelopes for flammable risks.
3) Terrain and obstacles

Flat desert terrain is not always “flat” in dispersion terms. Buildings, pipe racks, tank farms, and dunes create wakes and channelling that can trap or redirect clouds. For offshore platforms, superstructure and module layouts produce recirculation zones with persistent gas pockets if ventilation is restricted. Including building effects through empirical wake models or computational fluid dynamics (CFD) often determines whether a control room intake experiences a toxic excursion or not.
4) Modelling approaches and their fit-for-purpose use

Different methods exist along a spectrum of speed versus fidelity. The table below summarizes common approaches and where they are typically applied.
| Approach | Best Suited For | Strengths | Typical Pitfalls |
|---|---|---|---|
| Gaussian plume/puff | Neutrally buoyant gases over relatively open terrain; screening analyses | Fast; many regulatory precedents; easy sensitivity checks | Poor for dense gases, strong building effects, or complex terrain; limited for short-duration jets |
| Dense-gas integral models (e.g., SLAB-type formulations) | Heavier-than-air or cold releases (LNG vapor, hydrocarbons, chlorine) | Captures gravity slumping and layer growth; suitable for most onshore cases | Simplified obstacle treatment; may under/overpredict pooling in tight geometries |
| CFD (RANS/LES) | Complex building wakes, offshore modules, enclosed or semi-enclosed areas | Detailed spatial resolution; explicit obstacle representation; transient behavior | High input sensitivity; computationally intensive; requires careful validation |
Many studies adopt a tiered approach: start with integral/analytical models to narrow scenarios, then apply CFD to high-consequence locations such as air intakes or congested modules. This balances defensibility with practicality for project timelines.
5) Endpoints, criteria, and time horizons

Modelling should be anchored to relevant endpoints and exposure durations. For toxic effects, ERPG/AEGL thresholds or company-specific guidance define concentration and averaging times that matter for life safety and impairment. For flammable materials, LFL/UEL contours and thermal radiation from delayed ignition are key. Decision-makers must tie the endpoint to the use case — gas detector setpoints, control room integrity checks, fence-line risk, or muster safety — and maintain consistency across scenarios.
6) Data quality, uncertainty, and sensitivity
Uncertainty is intrinsic to atmospheric dispersion. Good methodology characterizes, rather than hides, that uncertainty. Useful practices include bounding cases (e.g., 1.5 m/s stable night vs. 5 m/s neutral day), and structured sensitivity on hole size, wind variability, and atmospheric stability. For select scenarios, comparing model outputs against historical experiments (e.g., neutrally buoyant Prairie Grass data or dense-gas field trials) helps contextualize the results. Internal peer review and model change control keep conclusions robust across project iterations.
For additional perspectives and safety insights curated within NPM’s ecosystem, professionals sometimes draw on knowledge sources like CABINET NPM, especially when aligning engineering judgments with broader governance and risk viewpoints. While not a substitute for site-specific modelling, such resources can help frame consistent assumptions across programs.
How This Works in Practice
Translating the technical framework into day-to-day practice depends on organizational rhythms, project phases, and the maturity of risk processes. A typical pathway builds on what has been established above and moves from definition to application.
- Scenario framing: Hazard studies (HAZID/HAZOP) and Bowtie workshops identify credible leaks and loss-of-containment drivers. These converge into a scenario register with frequencies (from failure data and LOPA) and consequence hypotheses (continuous vs. transient release, isolation times, blowdown assumptions).
- Data consolidation: P&IDs, cause-and-effect, blowdown logic, PSV sizing bases, and equipment lists inform source terms. Meteorological datasets are compiled with attention to representativeness for different seasons and diurnal patterns.
- Model selection: Screening models are scoped for broad coverage; CFD is earmarked for locations where building effects drive risk decisions (e.g., occupied buildings, offshore TRs, congested utility corridors).
- Execution and iterations: Analysts run cases, review outliers, and validate trends (e.g., denser clouds traveling further under stability F). They iterate with engineers when results suggest re-siting of detectors or intakes, or added barriers like deluge curtains.
- Integration into decisions: Results translate into gas detection layouts, ventilation specifications, control room performance standards, and emergency plans. They also feed QRA consequence modules, supporting individual and societal risk calculations.
- Documentation and submission: Reports capture assumptions, methods, inputs, and limitations in a way that aligns with ADNOC group HSE expectations, OSHAD SF documentation needs, and Civil Defence review practices.
Consider a few practical examples. In a sour gas facility, H2S clouds may remain dense and hug the ground during low-wind nights. A tiered modelling approach might first show potential for intake exceedances, triggering CFD around the control building to evaluate wake recirculation. In an ammonia terminal, releases through small-bore connections could produce transient plumes that challenge detector voting logic. In LNG processing, cold vapor clouds can pool in low-lying areas; integral models augmented with site-specific terrain inputs help visualize where curbing or drainage changes can reduce accumulation potential.
Offshore, even moderate jets can become trapped in module recirculation zones where mechanical ventilation is marginal, calling for CFD to predict accumulation times, detector response, and ESD/ventilation interplays. In urban-adjacent industrial zones, tank farms and warehouse rows create street-canyon effects; modelling should represent these obstacles realistically to set accurate emergency planning zones and support Dubai Municipality environmental submissions.
Benefits and Practical Use Cases
When anchored in sound physics and consistent assumptions, dispersion outputs enable targeted, effective decisions across the asset lifecycle. Some of the most frequent use cases in UAE industries include:
- Gas detection strategy: Setting detector densities, heights, and voting to catch both high-momentum jets and slow, dense clouds; coordinating with IEC 61511 safety function response times.
- Occupied building risk management: Verifying HVAC pressurization, filter choices, and air intake siting per API RP 752/753; defining shelter-in-place durations based on ERPG/AEGL averaging windows.
- Emergency planning and drills: Designing muster points, egress routes, and siren/gong zones; anticipating how sea breeze transitions may redirect hazard footprints over a shift.
- Layout and land-use planning: Evaluating separation distances for new units, setting boundary setbacks for flammable and toxic inventories, and supporting societal risk envelopes in QRA.
- Regulatory engagement: Substantiating submissions to Civil Defence, OSHAD, MOHRE-linked safety programs, and local environmental authorities with credible, reproducible modelling.
- Operational risk management: Assessing temporary operations (e.g., maintenance purges, nitrogen inerting, vent openings) and adapting permit-to-work controls to the predicted dispersion behavior.
Specific benefits follow naturally from these applications: fewer nuisance alarms and more meaningful detections; HVAC and building upgrades targeted to genuine vulnerabilities; emergency teams practicing scenarios that match meteorological realities; and more efficient stakeholder reviews thanks to transparent, auditable methods. Over time, organizations that institutionalize consistent dispersion methods reduce rework across projects and build a shared “mental model” for airborne hazards that spans engineering, operations, and leadership.
Best Practices and Points of Caution
Because dispersion outcomes drive high-stakes decisions, disciplined practices make the difference between a persuasive, defensible study and one that stalls in review.
- Make assumptions traceable: Tie every source-term parameter to a P&ID, datasheet, or study note (HAZOP/LOPA). Capture control logic that limits duration — it often matters more than hole size.
- Use representative meteorology: Build wind roses by stability class and season; consider sea-breeze cycles for coastal sites. Avoid cherry-picking “worst case” winds without context; present reasoned bounding cases with clear probabilities.
- Match model to physics: Use dense-gas models for cold/heavy releases; reserve CFD for obstacle-dominated problems or offshore modules. Document why the chosen method fits the question.
- Respect averaging times and endpoints: Toxic criteria depend on time windows; do not compare 1-minute peaks to 10-minute standards without conversion or caution.
- Represent obstacles credibly: Even “simple” sites have meaningful wakes. For occupied buildings and air intakes, treat obstacle impacts explicitly.
- Quantify uncertainty: Provide sensitivity runs and confidence commentary so decision-makers see robustness, not just single-value outputs.
- Integrate with broader safety processes: Feed results back into Bowties, LOPA targets, and SIS testing strategies. Misalignment creates gaps between paper and practice.
- Plan for change: As projects evolve, update critical scenarios under management of change (MOC). Small layout shifts can alter wake behavior and plume pathways.
- Communicate for use: Translate technical outputs into heat maps, contours, and simple narratives that ERT leaders and operators can act on under time pressure.
In UAE contexts, add two locale-specific cautions. First, night-time stability can dominate risk envelopes; schedule critical operations accordingly. Second, blending effects of high heat, humidity, and dust can influence plume buoyancy and sensor performance — adjust assumptions and detector technologies to keep predictions and field behavior aligned.
Frequently Asked Questions About Dispersion Modelling Methodology
How do I choose between a Gaussian model, a dense-gas model, and CFD?
Start with the physics of your release. If the gas is near-ambient temperature and density, with open terrain and modest obstacles, a Gaussian approach may be sufficient for screening. If it is heavier-than-air or cold (LNG vapor, some hydrocarbon mixtures, chlorine, ammonia), a dense-gas integral model is more appropriate. Use CFD where building wakes or module congestion dominate behavior, such as around control rooms or offshore platforms. Many teams apply a tiered approach: integral/analytical first, CFD second for critical hotspots.
What meteorological data should I use for a defensible study?
An onsite meteorological mast with at least 12 months of hourly data provides the best basis. Develop wind roses by stability class, and summarize diurnal cycles to capture sea-breeze patterns common in coastal UAE sites. If onsite data are not available, select the closest representative station and state the limitations. Align averaging times with toxic or flammable endpoints — do not mix 1-minute and 10-minute criteria without justification.
How are dispersion results typically integrated into a QRA?
Consequence modelling outputs feed into frequency-consequence pairs for individual and societal risk calculations. The dispersion module provides concentration-time footprints and ignition probabilities (where relevant) for each scenario. These results, weighted by scenario frequencies from failure data and LOPA, shape risk contours and potential loss-of-life estimates. Regulators such as ADNOC HSE often expect transparent linkage from HAZOP/LOPA scenarios to dispersion assumptions and then to QRA outputs.
What endpoints should I use for toxic releases?
Common practice uses ERPG or AEGL levels with averaging times that reflect incapacitation and escape objectives. Align endpoint selection with your corporate standards and local regulatory expectations, and make sure the modelling software uses the same averaging time. For building integrity studies, consider both threshold crossing and cumulative dose concepts to evaluate shelter-in-place durations and filter effectiveness.
Do I need CFD for every occupied building?
No. Use CFD selectively for locations where obstacle-driven recirculation or complex geometry meaningfully affects risk decisions — for example, a control building nestled among units with multiple nearby vents. For many other buildings, integral models with empirical wake corrections and conservative margins are sufficient. The key is to explain why CFD would or would not change the decision at hand.
How do dispersion results support gas detection design?
Models indicate where and at what heights gas clouds are likely to accumulate, guiding detector placement and density. They also inform expected concentration-time profiles, which help set alarm thresholds and voting logic consistent with IEC 61511 safety function response times. This reduces both false alarms and the risk of undetected hazardous accumulations.
What is different about offshore dispersion?
Offshore modules have denser equipment layouts and limited ventilation, which promote gas trapping and recirculation. CFD is often warranted to examine accumulation zones, detector coverage, and interactions with HVAC/ventilation shutdowns. The proximity of temporary refuges and the consequences of delayed ignition make detailed, obstacle-resolving modelling particularly valuable.
How should uncertainty be communicated to decision-makers?
Present bounding meteorological cases and key sensitivities (e.g., hole size, release duration, stability class), and show how recommended decisions hold across that range. Avoid implying false precision. Where feasible, compare model predictions to known experimental trends to ground the discussion.
Conclusion
Dispersion Modelling Methodology is ultimately about turning complex physics into clear, defendable decisions. By defining credible source terms, using representative meteorology, matching models to the physics of each scenario, and being explicit about uncertainty, organizations can align engineering choices, emergency plans, and regulatory submissions. In practice, this connects directly with the outputs of HAZOP, LOPA, and Bowtie analyses; informs QRA consequence layers; and guides tangible actions like detector placement, control room hardening, and muster strategy. Treated as a living discipline — updated through MOC, peer review, and lessons learned — it becomes a strategic asset for safety and performance.
To go further on this topic, NPM’s broader Toxic Dispersion Modelling Services UAE page covers this in more depth.
Explore NPM’s References
NPM’s track record spans complex consequence modelling and risk studies embedded in major capital projects and operational improvements worldwide. Our teams have supported refineries optimizing gas detection architectures after unit revamps, gas processing plants refining blowdown strategies to reduce toxic cloud persistence, and petrochemical complexes integrating occupied building risk management with ventilation and filtration upgrades. In logistics and storage terminals, we have used dense-gas models to evaluate low-temperature hydrocarbon releases across varied terrains, feeding into land-use planning and interface management with adjacent industrial tenants. Power generation facilities and utilities have drawn on our modelling to validate emergency planning zones and coordinate with municipal and civil defence authorities, building confidence that plans match likely plume behavior under different seasonal and diurnal patterns.
Across these contexts, NPM emphasizes transparency: tying assumptions to design data, making meteorological representativeness explicit, and selecting methods that balance realism with practicality. When building wakes or module congestion dominate, we deploy CFD with careful validation, while retaining integral models for broader scenario coverage. This disciplined approach enables stakeholders — engineers, ERT leaders, and reviewers — to align quickly on what the results mean and how to act on them. Whether the need is to support a QRA, refine a Bowtie barrier, or prepare a submission to authorities, NPM’s references show consistent delivery of credible, actionable modelling that integrates with standards such as ISO 45001, ISO 14001, API RP 752/753, and IEC 61511. The cases also reflect our fluency in working with regulatory frameworks and client governance structures, ensuring results travel smoothly from analysis to decision.
For more information about our references, visit Our References.
Speak With NPM
If you are considering how to apply dispersion modelling insights to a specific unit, building, or project decision, NPM is available to discuss practical approaches and typical pitfalls. We are happy to talk through data needs, model selection, and how results can be integrated into your existing risk processes without disruption. For any further questions or to speak with our team, Contact Us.