Empirical methods for determining elemental abundances tested on model H II regions.

On the ground of model H II regions of different chemical composition, exciting stars, and density distribution, it is shown that the temperature fluctuations inside nebulae with single exciting stars usually have less influence on the empirical determination of the elemental abundances than observational errors. However, for high overabundances (more than twice the solar heavy-element abundance) they may lead to underestimates by factors of 5-10. In the case of complex H II regions, the errors on the abundances arising from the use of a unique temperature determination are found to be about 30-50%; determination of the temperature of low-excitation regions would be desirable. An empirical rule for estimating the ionization correction factors for Ne and S is derived from the models. The correction factor for S is shown to be different from those used on the basis of ionization-potential considerations.