This study investigates the sensitivity of simulated emissions and airborne NO
and benzene concentrations to model input parameters. Emission factors were estimated using Mobile6.2 and ambient concentrations were estimated using AERMOD. The simulation was performed for Huron Church Road in Windsor, Ontario, Canada, an arterial road 9 km in length with an average vehicle speed of 50 km/h. Eight scenarios were developed such that one input parameter was changed at a time in the first seven scenarios and compared with the base case. Results showed that emission factors were most sensitive to the choice of vehicle composition (Ontario versus default), followed by the choice of vehicle age distribution (Ontario versus default), and the average speed of vehicles. Simulated concentrations were sensitive to the hour-of-day variation in emission (mainly due to variation in vehicle counts) and when this was not considered, the annual mean concentrations were likely overestimated by up to 27% and maximum hourly concentrations were underestimated. The findings provide insights into determining the level of details of input parameters required for estimation of emissions and concentrations.
Keywords: air pollution, mobile emissions, vehicle counts, diurnal variation, vehicle age distribution, dispersion, sensitivity