Generation of fuzzy rules from a European ozone model

This paper describes a pilot research aimed at the development of a simplified version of a European ground-level ozone model using a fuzzy rule generation methodology. The EMEP ozone model is used for predicting tropospheric ozone concentrations at every grid in Europe by taking into account the physical and chemical mechanisms. However, the EMEP model is too complicated for routine analysis of different policy options aimed at finding levels of reductions of emissions in each country that would result in decreasing the ozone concentration below given limits. Therefore there is a need to develop a simplified ozone model that can be verified by the EMEP ozone model and that can be used for rapid analysis of various policy options. The simplified ozone model should express the source-receptor relationships between ozone precursor emissions (NO/sub x/ and VOC) and ozone concentration in the troposphere, taking into account the meteorological conditions. One of the possible methodologies to develop such a simplified model is to use the fuzzy rule generation methodology. For some grids in Europe, fuzzy models are developed by using adaptive fuzzy clustering which gives a balance between linearity and continuity of the data distribution, adaptively.