Comparison of measured and model-calculated real-world traffic emissions

The quality of an emission calculation model based on emission factors measured on roller test stands and statistical traffic data was evaluated using source strengths and emission factors calculated from real-world exhaust gas concentration differences measured upwind and downwind of a motorway in southwest Germany. Gaseous and particulate emissions were taken into account. Detailed traffic census data were taken during the measurements. The results were compared with findings of similar studies. The main conclusion is the underestimation of CO and NO x source strengths by the model. On the average, it amounts to 23% in case of CO and 17% for NO x . The latter underestimation results from an undervaluation by 22% of NO x emission factors of heavy-duty vehicles (HDVs). There are significant differences between source strengths on working days and weekends because of the different traffic split between light-duty vehicles (LDVs) and HDVs. The mean emission factors of all vehicles from measurements are 1.08 g km −1 veh −1 for NO x and 2.62 g km −1 veh −1 for CO. The model calculations give 0.92 g km −1 veh −1 for NO x and 2.14 g km −1 veh −1 for CO. The source strengths of 21 non-methane hydrocarbon (NMHC) compounds quantified are underestimated by the model. The ratio between the measured and model-calculated emissions ranges from 1.3 to 2.1 for BTX and up to 21 for 16 other NMHCs. The reason for the differences is the insufficient knowledge of NMHC emissions of road traffic. Particulate matter emissions are dominated by ultra-fine particles in the 10–40 nm range. As far as aerosols larger than 29 nm are concerned, 1.80×10 14 particles km −1 veh −1 are determined for all vehicles, 1.22×10 14 particles km −1 veh −1 and an aerosol volume of 0.03 cm 3 km −1 veh −1 are measured for LDVs, and for HDVs 7.79×10 14 particles km −1 veh −1 and 0.41 cm 3 km −1 veh −1 are calculated. Traffic-induced turbulence has been identified to have a decisive influence on exhaust gas dispersion near the source.