This session aims to investigate sensitivity to parameterisations of process level studies in large scale prognostic models involving atmospheric aerosol particles. Aerosol schemes in large-scale models necessarily neglect the majority of known complex processes and will continue to do so in the future; so procedures for selecting the key processes of relevance to the climate system must be found. This, however, is difficult. For example, the computational expense of the complex representation of process level studies renders them inappropriate for large models, resulting in a trade-off between accuracy and efficiency. Similarly, despite the vast number of chemical species in the atmosphere, the chemical and physical properties and associated behaviour of the atmospheric aerosol, and appropriate precursors, are inadequately represented in state-of-the-science computer models aiming to predict climate change, greatly reducing our ability to attach any confidence to predicted outcomes.

With this session we wish to stimulate a debate on the methods available (or planned), and trade-offs made, for reducing the complexity of: aerosol thermodynamics; aerosol microphysics; and condensed-phase chemistry.