Primary versus secondary contributions to particle number concentrations in the European boundary layer

Abstract

Abstract. It is important to understand the relative contribution of primary and secondary particles to regional and global aerosol so that models can attribute aerosol radiative forcing to different sources. In large-scale models, there is considerable uncertainty associated with treatments of particle formation (nucleation) in the boundary layer (BL) and in the size distribution of emitted primary particles, leading to uncertainties in predicted cloud condensation nuclei (CCN) concentrations. Here we quantify how primary particle emissions and secondary particle formation influence size-resolved particle number concentrations in the BL using a global aerosol microphysics model and observations made during the May 2008 campaign of the European Integrated Project on Aerosol Cloud Climate Air Quality Interactions (EUCAARI). Observations are available from the DLR Falcon 20 aircraft and from 15 ground sites of the European Supersites for Atmospheric Aerosol Research (EUSAAR) and the German Ultrafine Aerosol Network (GUAN). Measurements include total and non-volatile particle number concentrations and the particle size distribution between ~3 nm and ~1 μm. We tested four different parameterisations for BL nucleation and two assumptions for the emission size distribution of anthropogenic and wildfire carbonaceous particles. When we emit small carbonaceous particles (recommended by the Aerosol Intercomparison project, AEROCOM), the spatial distributions of campaign-mean number concentrations >50 nm (N50) and >100 nm (N100) dry diameter were well captured by the model (R2~0.9) and the normalised mean bias (NMB) was also small (−5 % for N50 and 12 % for N100). Emission of larger particles, which we consider to be more realistic for global models, results in equally good correlation but larger bias (R2~0.8, NMB = −51 % and −21 %), which could be partly but not entirely compensated by BL nucleation. The model also predicts the particle concentration frequency distribution fairly well, with an overlap of modelled and observed N50 hourly histograms of ~60 % across all sites. However, the model-observation temporal correlation on an hourly time scale is poor (R2≤0.1) for this period. These comparisons show that caution is required when drawing conclusions about model realism from time or site-averaged data or frequency histograms when deterministic behaviour is not captured at individual sites. From this 1-month intensive European dataset it is not possible to determine a reliable estimate of the fraction of CCN-sized particles from primary and secondary sources, although the size of primary emitted particles is shown to be a major source of uncertainty.