Regional chemistry-transport models are used for atmospheric composition studies in several contexts : analysis of past events, scenarios studies, trends or forecast. Modelled concentrations are sensitive to many inputs data like anthropogenic surface emissions of NOx, VOCs and particulate matter. These emissions are provided as annual masses of pollutants for several activity sectors and projected onto a spatial grid. To use these data, modellers must make important assumptions in order to estimate pollutants fluxes for their own model grid and time frequency. Among these hypotheses, the time resolution is crucial and the way to redistribute emissions from annual to hourly fluxes determines the modelled concentrations accuracy. The usual CTMs approaches handle the time distribution with averaged factors. The present study quantifies with the chemistry-transport model CHIMERE the benefit of improving the calculation of traffic emissions fluxes by using hourly NO2 measurements nearby roadside areas as a proxy of road traffic sources. This work shows very different diurnal variation of emissions from country to country and suggests the use of a new hourly emission factor dataset for various countries. The induced changes are quantified for ozone, nitrogen dioxide and particulate matter surface concentrations over the whole Europe during the summer 2007. It is shown that the daily ozone peak remains relatively insensitive to this improvement whereas the pollutants concentrations during nighttime are closer to the measurements with the new profiles.