The York Research Database

This paper discusses the development of modelling techniques to understand radical chemistry in the troposphere. Data from recent field campaigns, combined with a detailed chemical mechanism, are used to construct box models to simulate radical concentrations. Comparisons with measurements are discussed. These mechanisms are typically of the order of 1700 reactions and 500 species, and are too large to gain an understanding of which reactions are driving the chemistry. Sensitivity analysis has been used to reduce the mechanisms. For reduced mechanisms that can predict the daytime concentrations of OH and HO2 to within 20% of the full mechanism, we can represent the chemistry of a remote coastal site in the marine boundary layer with only 25 reactions and 17 species. On a semi-polluted day, when the same site was also affected by isoprene from a local biogenic source, the chemistry can be represented by 64 reactions and 42 species. Algebraic steady-state solutions have been derived based on the reduced mechanisms. A local sensitivity analysis has been carried out to identify reactions and input concentrations that lead to the highest uncertainties in the model output. A global sensitivity analysis suggests that the 2 sigma uncertainty in model [OH] is +/- 42% and for model [HO2] is +/- 25%. (C) 2000 Elsevier Science Ltd. All rights reserved.