2004AJ....127.3484M

We examine the global accuracy of the commonly used multizone approach to inferring elemental abundances from nebular spectra. We use a large model grid of constant-density models to construct ``ionization correction factors'' (i_CF's) for several ions with bright optical lines. The emission lines from the model suite were analyzed as coming from a two-zone nebula, with temperatures and densities based on diagnostic line ratios. These were used to infer the ionic abundances responsible for the emission lines, to which the i_CF's were applied. The procedure was also applied to model grids simulating blister-type H II regions, as well as individual dense knots. Comparison of the model elemental abundances to those inferred from their line emission show large deviations with respect to the demonstrated accuracy of the analysis procedure; in addition, the inaccuracies of the three models follow significantly different trends, defeating any secondary correction appropriate for arbitrary nebulae. We thus strongly favor other methods for determining abundances from observations; e.g., direct modeling of photoionized nebulae or their small-scale structures.