A simulation approach to assess air pollution from road transport

A hybrid deterministic/statistical modeling approach is used to predict air pollution levels generated from motor vehicle emissions. The inclusion of the effects of source emissions and meteorology on ambient concentrations is allowed by the deterministic models, but only within their range of greatest reliability. Two deterministic models are used: the ATDL model to obtain an areal indication of pollution levels, and the GM model to provide information on concentration reduction with distance from the roadway. The output data obtained from the deterministic models can be extrapolated by using the statistical model to predict the maximum levels of pollution concentration. Monte Carlo simulation shows the effects of system uncertainty with respect to emissions. Variability in meteorology is taken into account by running the model separately with a time-series input of wind speed for different calendar years. The methodology is demonstrated with an example of an Australian town.