Trends in black carbon and size-resolved particle number concentrations and vehicle emission factors under real-world conditions
Kerbside concentrations of NOx, black carbon (BC), total number of particles (diameter > 4 nm) and number size distribution (28–410 nm) were measured at a busy street canyon in Stockholm in 2006 and 2013. Over this period, there was an important change in the vehicle fleet due to a strong dieselisation process of light-duty vehicles and technological improvement of vehicle engines. This study assesses the impact of these changes on ambient concentrations and particle emission factors (EF). EF were calculated by using a novel approach which combines the NOx tracer method with positive matrix factorisation (PMF) applied to particle number size distributions. NOx concentrations remained rather constant between these two years, whereas a large decrease in particle concentrations was observed, being on average 60% for BC, 50% for total particle number, and 53% for particles in the range 28–100 nm. The PMF analysis yielded three factors that were identified as contributions from gasoline vehicles, diesel fleet, and urban background. This separation allowed the calculation of the average vehicle EF for each particle metric per fuel type. In general, gasoline EF were lower than diesel EF, and EF for 2013 were lower than the ones derived for 2006. The EFBC decreased 77% for both gasoline and diesel fleets, whereas the particle number EF reduction was higher for the gasoline (79%) than for the diesel (37%) fleet. Our EF are consistent with results from other on-road studies, which reinforces that the proposed methodology is suitable for EF determination and to assess the effectiveness of policies implemented to reduce vehicle exhaust emissions. However, our EF are much higher than EF simulated with traffic emission models (HBEFA and COPERT) that are based on dynamometer measurements, except for EFBC for diesel vehicles. This finding suggests that the EF from the two leading models in Europe should be revised for BC (gasoline vehicles) and particle number (all vehicles), since they are used to compile national inventories for the road transportation sector and also to assess their associated health effects. Using the calculated kerbside EF, we estimated that the traffic emissions were lower in 2013 compared to 2006 with a 61% reduction for BC (due to decreases in both gasoline and diesel emissions), and 34–45% for particle number (reduction only in gasoline emissions). Limitations of the application of these EF to other studies are also discussed.