Optical nebular emission lines are commonly used to estimate the star formation rate of galaxies and the black hole accretion rate of their central active nucleus. The accuracy of the conversion from line strengths to physical properties depends upon the accuracy to which the lines can be corrected for dust attenuation. For studies of single galaxies with normal amounts of dust, most dust corrections result in the same derived properties within the errors. However, for statistical studies of populations of galaxies, or for studies of galaxies with higher dust contents such as might be found in some classes of transition galaxies, significant uncertainty arises from the dust attenuation correction. We compare the strength of the predominantly unobscured mid-IR [NeII]15.5um + [NeIII]12.8um emission lines to the optical H alpha emission lines in four samples of galaxies: (i) ordinary star forming galaxies, (ii) optically selected dusty galaxies, (iii) ULIRGs, (iv) Seyfert 2 galaxies. We show that a single dust attenuation curve applied to all samples can correct H alpha emission for dust attenuation to a factor better than 2. Similarly, we compare mid-IR [OIV] and optical [OIII] luminosities to find that [OIII] can be corrected to a factor better than 3. This shows that the total dust attenuation suffered by the AGN narrow line region is not significantly different to that suffered by the starforming HII regions in the galaxy. We provide explicit dust attenuation corrections, together with errors, for [OII], [OIII] and H alpha. The best-fit average attenuation curve is slightly greyer than the Milky-Way extinction law, indicating either that external galaxies have slightly different typical dust properties to the Milky Way, or that there is a significant contribution from scattering. Finally, we uncover an intriguing correlation between Silicate absorption and Balmer decrement.
Download