Predicting dust extinction from the stellar mass of a galaxy

TL;DR

This study finds that stellar mass is the primary factor determining dust extinction in star-forming galaxies, enabling a simple mass-based prediction model with broad applicability.

Contribution

It introduces a mass-dependent relationship to predict dust extinction, simplifying the correction process for galaxy observations across different wavelengths.

Findings

Dust extinction correlates strongly with stellar mass.

Residuals show minimal dependence on SFR or metallicity after mass correction.

A predictive relationship with ~0.3 mag scatter is established for H-alpha emission.

Abstract

We investigate how the typical dust extinction of H-alpha luminosity from a star-forming galaxy depends upon star formation rate (SFR), metallicity and stellar mass independently, using a sample of ~90,000 galaxies from Data Release 7 of the Sloan Digital Sky Survey (SDSS). We measure extinctions directly from the Balmer decrement of each source, and while higher values of extinction are associated with an increase in any of the three parameters, we demonstrate that the fundamental property that governs extinction is stellar mass. After this mass-dependent relationship is removed, there is very little systematic dependence of the residual extinctions with either SFR or metallicity, and no significant improvement is obtained from a more general parameterisation. In contrast to this, if either a SFR-dependent or metallicity-dependent extinction relationship is applied, the residual extinctions show significant trends that correlate with the other parameters. Using the SDSS data, we present a relationship to predict the median dust extinction of a sample of galaxies from its stellar mass, which has a scatter of ~0.3 mag. The relationship was calibrated for H-alpha emission, but can be more generally applied to radiation emitted at other wavelengths. These results have important applications for studies of high-redshift galaxies, where individual extinction measurements are hard to obtain but stellar mass estimates can be relatively easily estimated from long-wavelength data.