Systematics in the SED Fitting Parameter Estimation of Composite Galaxies

TL;DR

This study investigates biases in spectral energy distribution (SED) fitting of galaxy pairs, especially post-mergers, revealing minimal biases in stellar mass and star formation rate estimations despite differences in dust and star formation histories.

Contribution

It demonstrates that simple two-component star formation histories are sufficient for accurate SED fitting of unresolved galaxy pairs, providing a new empirical testing approach.

Findings

Biases in stellar mass and SFR are less than 0.1 dex for most galaxy pairs.

Moderate SFR biases (~0.1 dex) occur in pairs with contrasting dust content.

Simple two-component models are adequate for post-merger SED analysis.

Abstract

Derivation of physical properties of galaxies using spectral energy distribution (SED) fitting is a powerful method, but can suffer from various systematics arising from model assumptions. Previously, such biases were mostly studied in the context of individual galaxies. In this study, we investigate potential biases arising from performing the SED fitting on the combined light of two galaxies, as would be the case in post-merger systems. We use GALEX-SDSS-WISE Legacy Catalog (GSWLC) of z<0.3 galaxies to identify 9,000 galaxy pairs that could eventually merge. For these we investigate if the UV/optical SED fitting accurately determines the stellar mass and (specific) star formation rate if the pair was unresolved (merged). The sum of the stellar masses (and SFRs) of individual galaxies in the pair establishes the ground truth for these quantities. For star forming galaxies no biases (<0.1 dex) are found in the stellar mass, SFR, or sSFRs. Moderate systematics in SFR (~0.1 dex) are found for systems with an extreme contrast in dust content between the two galaxies. We conclude that biases that would arise in the determination of masses and SFRs of post-merger systems on account of the two original galaxies having potentially very different star formation histories and different dust properties are small and that the approach with simple two-component star formation histories is adequate. The approach presented in this study, using flux compositing with empirically determined ground truth, offers new opportunities for testing the results of SED fitting in general.