The Australian Air Quality Forecasting System (AAQFS) commenced operation in demonstration mode in August 2000, generating twice daily, numerical, 24-36 hour air quality forecasts for the states of Victoria and NSW. A range of air pollutants is forecast by the system, including ozone, nitrogen dioxide, carbon monoxide, sulfur dioxide, PM10 and PM2.5.

Ozone is one of the few pollutants that is observed to breach air quality objectives, and thus it is of particular interest to ascertain the ability of AAQFS to correctly forecast concentration extremes for this pollutant.

In this paper we will provide a brief description of the methodology used for forecasting photochemical oxidant. We will describe how this methodology evolved during and after the 2000/2001 photochemical smog season following feedback of system performance from specific case studies and from a consideration of routine daily performance metrics for that period. We will then go on to review the system performance for the 2001/2002 photochemical smog season- a period during which the methodology and system has remained relatively static. The performance of AAQFS will be contrasted with that of other contemporary systems.