| Popis: |
No dispersion model should be used without the associated errors in its predictions being presented. All model estimates should have ranges associated with them. The large number of parameters introduced into current dispersion models makes it difficult to estimate errors. This is a weakness of the models, not of the error analysis. Errors associated with different parts of a dispersion calculation should be distinguished. Errors associated with roads may be different to those for the urban background. The fuzzy approach has potential for use where one has separate well established, commonly used models e.g. CALINE4 for the road and OSPM for a street canyon, and one wishes to add the urban background using a well tested multi-source Gaussian model. The declaration of AQMAs in cities should be done in a way that takes account of the fuzziness in the setting of a AQMA. The AQMA can have sharp boundaries depending on choosing an α-cut of the underlying fuzzy set. The value of α chosen is a political decision depending on the degree of precaution selected. It is virtually impossible to display the declaration of an AQMA in a city on a map, because of the order of magnitude differences in the scale of the errors. |
