All models are wrong. Some are useful — George Box (Mathematician)
The Real Picture
Air quality modeling involves real mathematics, physics, and chemistry — there is no getting around that. And meteorological modeling, which feeds into it, carries its own complexity.
The physics that describes the atmosphere in one part of the world does not automatically translate to another. Localized emissions inventories, at the right (and high) resolution, are needed. A global inventory jumpstarts the work, but does not cut it.
That said, the situation today is genuinely different from what it was a decade back (aka 2010-2015). Computational resources are more accessible. YouTube has tutorials on all kinds of methods and models. Model-specific forums exist with full community users support. Setting up these systems is more within reach than it has ever been.
Now, air quality modeling demands commitment and it is achievable.
The “Monstrous” Reputation
In the low- and middle- income countries, where the health impacts are felt the most from increasing exposure to outdoor and indoor air pollution and where more information will only help us better understand the sources of air pollution, if modeling is more accessible than before, why are there not more groups running these exercises? Why are there not more training programs?
Part of the problem is how the conversation around modeling tends to start.
Before anyone has even touched a model or seen what it can produce, the discussion shifts to what is missing — what the model does not capture, what the limitations are, and what the complexities are.
In the atmospheric computational world outside of the US and the EU, we are operating in a data-poor environment where uncertainty is high. Gaps will always exist. Any model produced result will always leave someone unsatisfied. That includes the people building them.
But that is not a reason to start with the negativity and make the model the “monster”.
Hands-up or Learn
When someone explains a modeling methodology, uncertainty surfaces and that is natural.
There are two ways to respond to it. One is to ask how the calculation was done, because you want to learn and address any of the gaps. The other is to decide that it is too difficult and wait for a simpler, more complete method to arrive.
The reality is that a simpler method is not coming.
And in the meantime, the field stays small, the training programs stay rare, and groups stay uncomfortable talking about the numbers.
The Conversation
The focus has drifted toward problems — why models are not working, why they are not good enough. This framing scares people away before they even begin.
What this conversation needs instead is a focus on what is possible, on getting hands dirty with a model, and seeing what the model can bring to the table.
When uncertainty is part of the methodology explanation, it needs to be presented clearly, as an honest part of the process, and not as a warning sign.
The Change
The methods exist. The tools exists. The forums exist.
What is missing is operational trainings, the kind that builds confidence in the mathematics, the physics, and the chemistry of emissions and pollution modeling. The kind that makes groups comfortable talking about numbers, The kind that doesn’t stop by the idea that the numbers might not be perfect.
A short blog on “Why we are developing a course on air quality awareness and air quality management“.
Beginners handbook on assembling air pollution models [download]
Browse [link] for other FAQ’s