2.5 Modelling

Plume dispersion models are routinely used to inform assessments of air quality impacts, to either predict events or analyse past events. They can also be used in real time air quality management. The needs for real time models are different in some respects from those typically used for compliance assessments.

2.5.1 Basic description of models

Plume dispersion models mathematically simulate the dispersion (and deposition) of pollutants in the atmosphere after they are emitted from specific, defined emission sources. In common use today there are two main types of dispersion model: steady state and non-steady state.

Steady state models assume that, for each calculation of the plume (typically an hourly average), the meteorological conditions for that hour are in steady state—that is, they have always been and will always be the same. An example is the Australian regulatory model AUSPLUME, which is a steady-state Gaussian plume model, so named because it assumes that plume material, when averaged over time, has a Gaussian or normal distribution around the centreline of the plume. In performing its calculations, AUSPLUME steps from one hour to the next using the meteorological data for that hour to calculate the distribution of plume material downwind from the source(s). The steady state assumption means that each hour is independent of other hours.

A non-steady state model, on the other hand, tracks the location of plume segments through time. This means that variations in wind and other meteorological parameters that affect ground level concentrations can exert an influence on the predicted plume patterns.

2.5.2 Applications of modelling

Despite the sometimes large differences in model results for specific situations, the simpler steady state models such as AUSPLUME are widely used for regulatory purposes. However, more advanced, non-steady state models, such as TAPM (The Air Pollution Model) developed by CSIRO, and the CALPUFF model preferred by the United States Environmental Protection Agency, are increasingly being used as their costs and accessibility improve.

The main applications of dispersion models are:

  • predicting the impact of a proposed activity such as a mine or smelter
  • designing chimney heights or emission control systems
  • apportioning ambient impacts of emissions to specific sources
  • ranking emission sources in terms of priority for applying controls
  • analysing past air quality events
  • estimating emission rates (by back-calculation modelling—this can be particularly useful for estimating dust emissions from area sources).

Computational fluid dynamics (CFD) produces dynamic models that simulate the air flow over and around objects and barriers, or within enclosed and semi-enclosed spaces. CFD models have a multitude of applications, ranging from assisting in the design of combustion chambers and ventilation systems, to designing barriers to prevent dust lift-off. With respect to dust, they are particularly useful in understanding how far dust will be carried, where it will land and how effective the wind barrier will be.

Examples of CFD air flow lines modelled using computational fluid dynamics. Source: Richard Meloy, Rio Tinto.
Examples of CFD air flow lines modelled using computational fluid dynamics. Source: Richard Meloy, Rio Tinto.

2.5.3 Modelling cumulative impacts

For the assessment of a new mine or emission source, the existing levels of dust (or other pollutants as relevant) need to be taken into account. For pollutants which have well-defined sources, such as sulphur dioxide, it is possible to include the new sources as well as existing or background sources in the model and yield acceptable results. However, for particles, a complete accounting of background sources is not possible: the background of airborne particles comes from a variety of local and distant activities such as natural wind erosion, agriculture, industry and transport.

In the case of PM10, for example, the best approach is to conduct monitoring and toapply the results as a background to which the new sources are added. Depending on the sensitivity of the activity and regulatory needs, the background can be included either as a fixed value, such as the seventieth percentile of the daily values, or as daily or hourly varying background. However, particularly where there are multiple mines or other sources nearby, such as in the Hunter Valley, cumulative impact modelling is not straightforward and there are multiple sources of uncertainty.

2.5.4 Model validation and uncertainty

All models are simplifications of the real world and carry inherent uncertainty, as well as uncertainty associated with inaccuracies in the input data. Data on emissions used in an impact assessment may turn out to be significantly in error; particularly if the assessment shows a small difference between predicted impacts and the acceptable limits, the agency approving a mining proposal may require the model to be validated.

Once the operation is underway, this involves monitoring ambient PM10 (or another critical air quality indicator) and meteorology for a year or two, then compiling a more accurate emissions inventory. With the new data, the model is re-run and the results are compared to the measured PM10. Once a validated model exists, any future expansion or changes to emissions can be predicted with greater confidence.

Share this Page