Parameterising secondary organic aerosol from α-pinene using a detailed oxidation and aerosol formation model

A new parameter model for alpha-pinene secondary organic aerosol (SOA) is presented, based on simulations with the detailed model BOREAM (Biogenic hydrocarbon Oxidation and Related Aerosol formation Model). The parameterisation takes into account the influence of temperature, type of oxidant, NOx-regime, photochemical ageing and water uptake, and is suitable for use in global chemistry transport models. BOREAM is validated against recent photooxidation smog chamber experiments, for which it reproduces SOA yields to within a factor of 2 in most cases. In the simple chemical mechanism of the parameter model, oxidation of alpha-pinene generates peroxy radicals, which, upon reaction with NO or HO2, yield products corresponding to high or low-NOx conditions, respectively. The model parameters - i.e. the temperature-dependent stoichiometric coefficients and partitioning coefficients of 10 semi-volatile products - are obtained from simulations with BOREAM, including a prescribed diurnal cycle for the radiation, oxidant and emission levels, as well as a deposition sink for the particulate and gaseous products. The effects of photooxidative ageing are implicitly included in the parameterisation, since it is based on near-equilibrium SOA concentrations, obtained through simulations of a two-week period. In order to mimic the full BOREAM model results both during SOA build-up and when SOA has reached an equilibrium concentration, the revolatilisation of condensable products due to photochemical processes is taken into account through a fitted pseudo-photolysis reaction of the lumped semi-volatile products. Modelled SOA mass yields are about ten times higher in low-NOx than in high-NOx conditions, with yields of more than 50% in the low-NOx OH-initiated oxidation of alpha-pinene, considerably more than in previous parameterisations based on smog chamber experiments. Sensitivity calculations indicate that discrepancies between the full model and the parameterisation due to variations in assumed oxidant levels are limited, but that changes in the radiation levels can lead to larger deviations. Photolysis of species in the particulate phase is found to strongly reduce SOA yields in the full model. Simulations of ambient conditions at 17 different sites (using oxidant, radiation and meteorological data from a global chemistry-transport model) show that overall, the parameterisation displays only little bias (2%) compared with the full model, whereas averaged relative deviations amount to about 11%. Water uptake is parameterised using fitted activity coefficients, resulting in a good agreement with the full model.