De l’intérêt de contrôler l’impact des hypothèses de composition du parc automobile sur l’estimation des émissions liées au trafic routier

Since the 1990s, transport project assessments take systematically pollutant emission estimations into account. This paper is about the methodological aspects of these calculations. It focuses more specifically on the car fleet hypothesis, which most often lays on national data, without consideration of local specificities. We use the last household travel survey from Lyon, 2006, and the SIMBAD model to compare the results of CO 2 and NO x emissions estimated from the French national car fleet, the aggregated Lyon car fleet and the same fleet disaggregated by household location and income. We show that the error level varies, depending on the pollutant and the observation scale. The use of an aggregated local car fleet seems interesting and satisfactory for a global emission assessment. If the results are required at a more detailed spatial level, the use of this local fleet improves sharply the estimations in comparison of a national fleet; the fleet disaggregation refines the results for NO x .