A model is a simplified representation of reality. We do not transfer every detail of nature into a computer. We select the factors most relevant to a particular process and represent their relationships to help us understand or predict it. That is why I introduce five elements: the variable of interest, external influences, parameters describing the system, constants and the equations connecting them.
First: the variable we want to calculate
A variable is the quantity whose condition or variation in space or time interests us. In dispersion modelling, we might want SO₂, NO₂ or PM10 concentration 500 metres from a source, at the nearest residential area or along the project boundary. The question concerns what reaches those locations, rather than only what leaves the stack.
This leads to further questions: where is the highest concentration, when does it occur, what averaging period applies and can the result be compared with the chosen assessment basis? The purpose is to understand conditions at receptors and discuss their implications, not merely produce a map.
Second: the external conditions
Meteorology provides clear examples: wind speed and direction, temperature, atmospheric stability and mixing height. The same source at the same emission rate can produce different results under different conditions, because it releases into an environment rather than an empty space.
Strong and weak winds differ, as do winds directed towards a residential receptor and winds in another direction. Terrain and surrounding surface characteristics also contribute to the represented environment, depending on the model. These are part of the calculation, not background facts to mention once and then ignore.
Third: parameters describing the source and system
Stack height, diameter, gas exit velocity and temperature help describe a point source. Some values may remain fixed within a defined operating scenario and change when another design is tested. In that comparison we have changed the source description rather than the weather.
Calling something a parameter does not mean it must always be physically constant. Operating conditions can vary, and the available evidence and methodology may support time-varying representation. Distinguish what is held fixed from what changes, and make sure the values and units describe the actual scenario.
Fourth: constants and reference quantities
Formulations use physical constants or reference values, such as an approximate gravitational acceleration of g = 9.81 m/s² in appropriate applications. Their role differs from a stack height selected by a designer or a wind speed recorded by an instrument.
Recognising that difference helps establish which quantities may legitimately change between scenarios and which follow the model definition. It also reinforces the need for consistent units: a familiar number is useful only when supplied in the form and units the calculation expects.
Fifth: the equations connecting the evidence
Equations turn the description of source and environment into the required output. Some air dispersion applications use Gaussian formulations, while other models use different approaches. One formulation is not automatically appropriate for every source, distance or atmospheric situation.
A beginner does not need to memorise every equation before starting, but should understand the scientific basis and applicability. The result depends on inputs and assumptions. If those are incorrect or outside the model’s valid application, precise-looking numbers and attractive graphics do not repair the problem.
Using the framework to review a result
When an output changes, I ask whether the requested variable or averaging period changed, whether meteorology changed, whether a source property was altered or whether a calculation setting differed. This makes review a search for an explainable cause rather than random adjustments aimed at obtaining a lower number.
The five elements are a teaching framework, not a rigid taxonomy imposing identical terminology on every discipline. They organise understanding before model selection. Knowing each component’s role makes it easier to discuss both the conclusions and the limits of the calculation.
References & original writing
Adapted and expanded from the author’s LinkedIn writing. Technical references are listed below.
