This paper presents the results of simulations performed with CAMx v6.40 over the Po Valley area (Northern Italy), aimed at improving the prediction of organic aerosol (OA) levels and acquiring information on the sensitivity of CAMx with respect to different input configuration characteristics, in particular: (I) volatility distributions of organic emissions, (ii) parameterization of emissions of semi- and intermediate-volatile compounds (S / IVOC), and (iii) aging processes. These analyses are based on the latest experimental information available in the scientific literature. Model results were validated against two OA-specific datasets available for both an urban site (Bologna, February 2013) and rural site (ISPRA, March 2013). A significant improvement was observed in the model performance during reconstruction of OA levels by applying new S/IVOC emission parameterizations along with new volatility distributions, at both validation sites.

These results show that organic particulate emissions in the semi-volatile and intermediate-volatile range are most likely underestimated in official emission inventories, for each major source category (e.g., biomass combustion, diesel and gasoline vehicle exhaust). Finally, the model results showed low sensitivity to ageing processes, due to the low photochemical activity generally observed in winter. However, the study showed that the activation of SOA aging processes due to biomass burning, which is disabled by default in CAMx v6.40, may help bridge the gap between modelled and observed SOA concentrations.