A GIS approach to modelling traffic related air pollution

There is increased concern regarding the effect of traffic related pollution on public
heath. As the number of vehicles on the roads continues to rise, it is becoming
increasingly more important to identify areas where the population may be at a greater
risk to raised levels of pollution and areas where the implementation of policy to
control and monitor levels of pollution would be beneficial.

Traditionally, levels of air pollution have been established through dispersion modelling
or monitoring. However, for modelling traffic related pollution for large populations,
these methods have proved inappropriate.

Three new approaches have been developed to model traffic related air pollution and
are reported in this thesis. The approaches have been developed in a Geographical
Information System (GIS) and involve generating detailed maps of the pollution
surface from monitored data and information about the pollution sources. The new
methods are compared against the geostatistical technique kriging.

The first approach combines spatial interpolation from monitoring sites and dispersion
modelling, linking the dispersion model to the GIS, the second combines GIS
techniques for filtering data and spatial interpolation, and the third uses a combination
of GIS techniques for filtering and statistical techniques.

The three approaches are tested and validated by predicting levels of pollution at
monitoring sites not used to develop the models. It was found that the new
approaches provided more reliable estimates of pollution at unsampled locations than
kriging, with the last of these proving to be the most effective. The adjusted r2 values
for kriging, interpolation and dispersion, interpolation and filtering, and filtering and
statistics were found to be 0.44, 0.63, 0.67 and 0.82 respectively.

The approaches therefore have clear potential in the areas of air pollution management
and epidemiology, where the maps can be used to help identify locations where levels
of pollution exceed air quality standards, assess the relationship between air pollution
and health outcome and examine the risk of exposure to raised levels of pollution.