Galaxy properties from J-PAS narrow-band photometry

TL;DR

This study compares galaxy property estimations from narrow-band photometry and high-resolution spectroscopy, finding that narrow-band data provides similar trends but with less precision, and highlighting biases in Bayesian analysis methods.

Contribution

It demonstrates that narrow-band photometry can reliably recover galaxy properties with comparable trends to spectroscopy, while analyzing biases introduced by prior assumptions in Bayesian SED fitting.

Findings

Narrow-band photometry yields similar galaxy property trends as spectroscopy.

Spectroscopy provides approximately twice the precision of narrow-band photometry.

Prior PDFs in Bayesian methods can introduce biases not accounted for in the analysis.

Abstract

We study the consistency of the physical properties of galaxies retrieved from SED-fitting as a function of spectral resolution and signal-to-noise ratio (SNR). Using a selection of physically motivated star formation histories, we set up a control sample of mock galaxy spectra representing observations of the local universe in high-resolution spectroscopy, and in 56 narrow-band and 5 broad-band photometry. We fit the SEDs at these spectral resolutions and compute their corresponding the stellar mass, the mass- and luminosity-weighted age and metallicity, and the dust extinction. We study the biases, correlations, and degeneracies affecting the retrieved parameters and explore the r\^ole of the spectral resolution and the SNR in regulating these degeneracies. We find that narrow-band photometry and spectroscopy yield similar trends in the physical properties derived, the former being considerably more precise. Using a galaxy sample from the SDSS, we compare more realistically the results obtained from high-resolution and narrow-band SEDs (synthesized from the same SDSS spectra) following the same spectral fitting procedures. We use results from the literature as a benchmark to our spectroscopic estimates and show that the prior PDFs, commonly adopted in parametric methods, may introduce biases not accounted for in a Bayesian framework. We conclude that narrow-band photometry yields the same trend in the age-metallicity relation in the literature, provided it is affected by the same biases as spectroscopy; albeit the precision achieved with the latter is generally twice as large as with the narrow-band, at SNR values typical of the different kinds of data.