Air Quality Modeling: From Emissions to a Decision

What does air quality modeling actually tell you about an industrial project?

Before an industrial facility is built, expanded or modified, someone has to answer a deceptively simple question:

Where will the emissions go, how high will concentrations become, and will the project meet the applicable air-quality requirements?

Modeling dispersion can help us prevent air pollution

That is what air-quality dispersion modelling is designed to determine.

A dispersion model combines information about the facility, emissions, buildings, terrain and weather to estimate concentrations at locations where people or environmental receptors could be affected.

But running the model is only part of the job.

The difficult part is deciding:

what should be modelled,
how it should be represented,
whether the assumptions are realistic and
whether the results make physical sense.

That is where professional air quality modeling experience matters.

The basic idea is simple

An industrial source releases a contaminant. The atmosphere transports and disperses it. The concentration changes with:

emission rate
stack height
exhaust temperature
exhaust velocity
source location
building dimensions
terrain
wind speed and direction
atmospheric stability
mixing height
atmospheric chemistry
deposition
background concentrations

The model then predicts concentrations at selected receptors.

Those predictions can be compared with applicable air-quality objectives, standards, guidelines or health-based benchmarks.

In Alberta, air quality modeling guidance and tools are intended to help practitioners select appropriate modelling approaches and identify data requirements. Alberta's Modelling Expert System, for example, is specifically designed to help modellers determine what modelling tools and data are appropriate for a particular objective.

Coloured drawing of a production facility

So the real workflow is:

Facility → Emissions → Meteorology → Dispersion → Concentrations → Regulatory criteria → Decision

The computer performs the calculations.

The modeller has to make the decisions.

What does an air-quality modeller actually do?

A good assessment starts long before anyone presses "Run." Understand the facility.

The first step is understanding what is actually being built or operated. That means reviewing information such as:

facility layout
process descriptions
emission sources
stack locations
stack heights
stack diameters
exhaust temperatures
exhaust velocities or flow rates
operating schedules
production rates
fuel types
control equipment
buildings and structures
terrain
nearby industrial sources

This sounds straightforward. It isn't.

Engineering drawings describe what a facility is supposed to be. A dispersion model needs to represent what the facility can actually do.

That distinction can matter.

What an Experienced Modeller Notices that a Generic Consultant may Miss

This is one of the most important parts of the job. A modeller can receive a perfectly formatted emissions spreadsheet and still have the wrong model.

Why? Because the numbers may not represent a physically realistic operating scenario. For example:

Can all the sources actually operate at their maximum rates simultaneously?

A spreadsheet may say yes. The process may say no.

Does the stack parameter make physical sense?

A small change in exhaust temperature or flow can alter plume rise and therefore ground-level concentrations.

Is a source really continuous?

A source that operates for ten minutes during a particular process event may need to be treated very differently from a source operating continuously.

Is the building configuration correct?

Buildings can alter the flow around a stack and cause building downwash, potentially increasing concentrations near the facility.

Is the meteorology representative?

Five years of data (which we normally use in Alberta) can produce thousands of modelled hours, but the number of hours is not what makes a meteorological dataset appropriate.

The question is whether the dataset reasonably represents the conditions relevant to the project.

Does the maximum concentration make physical sense?

A model can produce a mathematically valid maximum that deserves a second look.

The model output is not automatically the answer simply because the software produced it.

That is one of the distinctions between operating a dispersion model and practising air quality modeling.

A practical air quality modeling workflow

A typical assessment can be thought of as nine questions.

Question

What needs to be established

What is being emitted?

Contaminants, emission rates and source characteristics

Where is it emitted?

Stacks, vents, tanks, fugitives and other sources

When is it emitted?

Normal, maximum, startup, shutdown and abnormal conditions

How does it enter the atmosphere?

Flow, temperature, velocity, release height and source geometry

How will the atmosphere transport it?

Meteorology, terrain, stability and mixing

What happens after release?

Dispersion, chemical transformation and possibly deposition

Where could impacts occur?

Receptors, property boundaries, residences and sensitive locations

What concentrations matter?

Applicable regulatory or health-based criteria

Does the result make sense?

Physical interpretation, sensitivity testing and professional review

That last question is particularly important. Does the result make physical sense? A model should never be treated as a black box.

Choosing the Right Model

There is no prize for using the most complicated model. The appropriate model depends on the regulatory question and the characteristics of the project.

Common modelling systems include AERMOD and CALPUFF, among others.

Alberta's modelling guidance recognizes that unusual combinations of terrain, climate, source configuration, emissions characteristics, receptor sensitivity or other circumstances can require a modelling approach better suited to the specific problem.

For example, Alberta guidance describes AERMOD as a steady-state plume model and CALPUFF as a non-steady-state puff model capable of representing time- and space-varying meteorological conditions.

The important point is not: Which model is best?

It is: Which modelling approach best answers this particular question? That distinction can save considerable time and money.

What can change the answer?

One of the most useful things a modeller can do is identify the variables that actually drive the result. For a stack source, those might include:

Emission rate → How much contaminant is released?

Stack height → How far above ground is the release?

Exhaust temperature → How much plume rise occurs?

Exhaust velocity → How much momentum does the plume have?

Building configuration → Does downwash bring the plume toward the ground? And where?

Meteorology → Under what conditions does the plume disperse?

Terrain → Does the landscape redirect or alter the plume?

Operating scenario → What sources are actually operating together?

Changing one of these assumptions can change the predicted maximum concentration. That is why model review should focus on sensitivity, not simply the final number.

A model can be Technically Correct and still Answer the Wrong Question

This is perhaps the most expensive modelling mistake.

Imagine a facility expansion. The existing air-quality assessment is still valid for the original plant.

The company adds several new emission sources. Someone proposes simply rerunning the old model. But what has changed?

New sources may have been added.
Operating scenarios may have changed.
Buildings may have changed.
Stack parameters may have changed.
Production rates may have changed.
The relevant receptors may have changed.
The regulatory question may have changed.

The old model may still run perfectly. It may simply no longer represent the facility being assessed. A technically flawless answer to the wrong question is still the wrong answer.

When should you consider air-quality modelling?

You may need a new assessment, or a review of an existing one, when you are:

building a new facility
increasing production
modifying a stack
adding combustion equipment
installing or removing a flare
changing fuel
changing operating conditions
adding or modifying emission controls
applying for an approval amendment
seeking temporary authorization
responding to a regulator
evaluating cumulative effects
determining whether an existing assessment remains valid

But we changed something does not automatically mean we need a completely new model.

Sometimes an existing assessment already answers the question. Sometimes only the emissions inventory needs to be updated.

Sometimes the important issue is a source parameter. Sometimes a sensitivity run is enough. Sometimes the air quality modeling approach itself needs to change.

The first question should therefore be: What does the regulator actually need demonstrated?

From Modelling Problem to Engineering Solution

Air-quality modeling shouldn't simply identify a problem and hand it back to the engineering team. The useful part comes next.

If predicted concentrations are higher than desired, potential solutions might include:

modifying stack height
changing exhaust conditions
modifying operating scenarios
relocating a source
changing building configuration
improving emission controls
changing fuel
modifying flare operation
reducing an emission rate
changing the timing of an operation

Then the proposed solution can be modelled again. The objective is not necessarily to make the model produce a lower number. The objective is to find a practical facility design or operating condition that works in the real world.

Air quality modeling for an exhaust stack

Model output in 3D

What Experience Looks like in Practice

BARRY J. LOUGH, EP - DISPERSION METEOROLOGIST

My experience with dispersion modelling began with meteorology and physics rather than with an air quality modeling software package.

I earned a B.Sc. specializing in Physics from the University of Alberta and an undergraduate Science Diploma in Meteorology. My work has included meteorology, dispersion modelling, emissions inventories, air monitoring analysis, regulatory assessments, data validation, operational meteorology and environmental problem-solving.

Over the years, I have worked on projects involving:

upstream and midstream energy facilities
chemical and petrochemical facilities
oil storage facilities
cogeneration and power facilities
fertilizer plants
refineries
pipeline projects
large multi-facility industrial complexes
unusual emission sources
cumulative industrial emissions

The work has sometimes required going well beyond a standard model run.

For a detailed, Myers Briggs assisted, description of my personal preferences, see this page.

A few Examples of the Problems I've Actually Worked on

Joffre Industrial Complex

At the Joffre Alberta Industrial Complex, modelling involved multiple industrial facilities operating in close proximity.

The work required detailed information about source locations, emissions, operating conditions and building effects.

Importantly, the air quality modeling team had to work with operational personnel to understand how the facilities actually operated.

That is a recurring theme in industrial modelling: The most important information is not always in the engineering drawings.

Operators often know which equipment can run simultaneously, which sources are intermittent and which operating conditions are realistic.

That knowledge can materially improve an assessment.

Hardisty Tank Storage Complex

Modelling work at the Hardisty Tank Storage Complex involved a large industrial setting with numerous tanks and associated infrastructure.

Projects like this illustrate another important modelling issue: The facility being assessed may not be the only facility contributing to the predicted concentration.

Nearby industrial sources can contribute to cumulative impacts.

Refinery and petrochemical projects

My experience has also included modelling emissions from large industrial complexes involving contaminants such as:

NOₓ
NO₂
SO₂
H₂S
CO
VOCs
benzene
butadiene
particulate matter
ammonia
other process-specific contaminants

Some projects required consideration of atmospheric chemistry, including the conversion of NOₓ to NO₂. Others involved unusual sources or specialized operating conditions.

Sometimes the Interesting Problem isn't Pollution

One project involved predicting fog conditions near roads adjacent to power and cogeneration facilities.

Potential Futuristic Industrial Facility

Future industrial facility?

Water vapour can become an operational issue when released under the right meteorological conditions.

This is a useful reminder that atmospheric modelling isn't simply about asking: Where does the pollution go?

It can also answer: What will the atmosphere do with what the facility releases?

That distinction becomes especially important when meteorology is part of the operational problem.

What I Look for when Reviewing an Existing Model

Before accepting a dispersion modelling assessment, I want to know:

1. Are the emissions realistic?

Not merely plausible on paper. Can the facility actually operate this way?

2. Are the sources represented correctly?

Are stacks, vents, tanks, fugitives, buildings and other relevant sources included?

3. Are the operating scenarios defensible?

Does the model represent realistic combinations of equipment and operating conditions?

4. Is the meteorology appropriate?

Does it represent the site and the conditions relevant to the assessment?

5. Is the model appropriate for the problem?

A familiar model is not necessarily the appropriate model.

6. Are the receptors appropriate?

Where are people, residences, property boundaries and environmental receptors?

7. What drives the maximum?

Can we explain why the maximum concentration occurs where and when it does?

8. Does the result make physical sense?

This is where professional judgement becomes particularly important.

9. What happens if the important assumptions change?

Sensitivity testing can reveal whether the conclusion is robust—or whether it depends on one questionable assumption.

10. Does the assessment answer the regulator's actual question?

If not, more model runs may simply produce more irrelevant information.

The Modeller's Job is to Reduce Uncertainty. A dispersion model does not give us a perfect picture of the future. It gives us a structured way to evaluate what could happen under defined conditions.

The quality of the answer depends on the quality of:

the emissions + the facility representation + the meteorology + the model + the assumptions + the interpretation.

That is why good modelling is more than a software exercise. It is an exercise in physics, meteorology, engineering, environmental regulation and professional judgement.

Can you do it better?

You like this Idea? It works for its intended purpose, but probably is not perfect. Why not?

Do you know of a better one? We'd all like to see it, I'll bet. A penny for your thoughts.

What shall we call it?

And What's it all About? [ ? ]

If you've got a picture...[ ? ]

Add a Picture/Graphic Caption (optional)

Click here to upload more images (optional)

Author Information (optional)

To receive credit as the author, enter your information below.

Your Name	(first or full name)
Your Location	(e.g., City, State, Country)

Submit Your Contribution

Check box to agree to these submission guidelines.
I am at least 16 years of age.
I understand and accept the privacy policy.
I understand that you will display my submission on your website.

(You can preview and edit on the next page)

Other people and their ideas...

Click below to see hi-tech contributions from other visitors to this page...

Control pollution and save planet
Today we are in a situation that we cant live without comfortaility. But for comfort we are destroying our nature and envirounment. This will leads to …

Air Pollution Modelling
First, I would like to intruduce myself. I am a young (26) chemical technology engineer, very interested in air pollution modelling. The reason for …

air pollution dispersion modeling
Learn how to model air quality with AERMOD and CALPUFF! Students and researchers can use these powerful modelling tools thanks to EPA software like …

Save your life... Save Envinronment.... Reduce Pollution
Air pollution is a point of major concern now. Statistics shows that there is an enormous increase in breathing and chest related diseases in urban areas. …

Regional air quality
I and everyone are concerned about regional air quality. It's a basic need for our future generation. The world already has a lot of pollution. …

Air Quality Modelling Update for 2009
Alberta Environment (AENV) has had a few official names over the years. Since 1971, this ministry has gone through a lot of revisions, integrations, and …

Clean Up the Air
We see the need to clean up the air. Regional air quality is a major concern, especially in the Detroit-Windsor region. My husband and son have asthma. …

Need for polluton control
Today we are in a situation where everyone needs comfort-ability. Due to this comfort all destroying the environment and the nature. This could lead …

Drive a Hybrid
It's called stop buying gas guzzling machines! Barry's Response - If you're not familiar with the term, a hybrid electric car uses a regular engine …

Environment is degrading
Now-a-days many issues regarding atmosphere research is going on in order to come out of the pollution. From the moment when we come out of the house …

Air Quality Dispersion Models
Do you know where I can find air quality dispersion models? Hi....im an environmental engineering student,looking for some data about air pollution …

Air Modelling
Q: I need help with air modelling. I want to model air pollution at a point surrounded by several stacks. What will I do and how? Please help..... …

computer models
1. Currently there is no scale existing to test the accuracy of any EPA computer models for prediction of Air Pollution, any where in the world. In a recent …

Modelling uncertainty and errors
Here is what the EPA has to say about modelling uncertainty and errors. It incorporates some theory put forth by Carl Gauss: Modeling software is more …

Click here to write your own.

What does a Good Air Quality Modeling Assessment Give you?

At the end of a project, the useful product isn't a stack of model output files. It is an answer.

Will the project meet the applicable air-quality requirements?

And, if there is a problem:

What is causing it?
How sensitive is the result?
What can be changed?
Does the proposed solution work?
Can the conclusion be defended to the regulator?

Those are the questions that matter to project managers, engineers, environmental professionals and regulators.

Air-quality modelling is ultimately about making better decisions. I've spent much of my career working at the point where meteorology, engineering and environmental regulation meet.

That experience has taught me that the most valuable modelling work often happens before the model is run—and after the computer has finished calculating.

Before the run, you have to determine what the facility really does. After the run, you have to determine whether the results make physical sense. Between those two steps is where professional judgement earns its keep.

If you are considering a new project, modifying an existing facility, responding to a regulatory question or wondering whether an existing assessment is still valid, start with the question—not the software.

Contact Calvin Consulting Group Ltd. about your air-quality modelling requirements

Clean air is our Passion...Regulatory Compliance is our Business.

Want to go Deeper?

Air-quality modelling sits at the intersection of meteorology, physics, engineering and environmental regulation. If you want to understand why a model produces the results it does, the weather is a good place to start.

Explore:

How atmospheric stability affects dispersion - Why stable air can trap pollutants while unstable air promotes vertical mixing.
How temperature inversions work - One of the most important meteorological conditions affecting near-surface air quality.
How wind affects pollution dispersion - Why wind speed and direction can matter as much as the emission rate.
How clouds form from atmospheric stability - A useful introduction to the same atmospheric physics that influences dispersion.
AERMOD vs. CALPUFF - When the two modelling approaches differ and why model selection matters.
How to check dispersion-modelling data quality - A practical look at the assumptions and inputs that should be examined before trusting model output.
Air quality modeling and monitoring - Why measurements and models answer different questions and why they can be much more useful together.

The more you understand the atmosphere, the easier it becomes to understand what a dispersion model is actually telling you.

To see articles and notes about air issues on a regular basis, read them in Blowin' in the Wind

Keyword for this page: air quality modeling

Search this site for more information now.

Go back from Air Quality Modelling to the Air Quality Testers web page or visit the Stuff in the Air homepage.

Share this page:

What’s this?

Facebook X

Know someone who needs to see this. Here's an easy way show them.

Would you prefer to share this page with others by linking to it?

Click on the HTML link code below.
Copy and paste it, adding a note of your own, into your blog, a Web page, forums, a blog comment, your Facebook account or anywhere that someone would find this page valuable.

var l = window.location.href, d = document; document.write('<form action="#"><div style="text-align:center"><textarea cols="50" rows="2" onclick="this.select();"><a href="'+l+'">'+d.title+'</a></textarea></div></form>');