Speaker
Description
Type Ia supernovae remain one of the key observational probes of dark energy, but their cosmological constraining power depends on constructing Hubble diagrams with well-understood biases. In future high-statistics time-domain surveys, including TiDES and LSST-era spectroscopic follow-up programmes, host-galaxy redshifts will often need to be obtained under finite spectroscopic resources. It is therefore important to understand whether shallow host-galaxy follow-up recovers all galaxy populations equally, and whether any missed hosts could introduce biases into supernova cosmology.
Using OzDES DR2 spectra together with DES-SN5YR host-galaxy data, we construct shallow three-exposure spectra for 1,280 supernova host galaxies and test redshift recovery using MARZ. We find a strong host-dependent selection effect: star-forming hosts are recovered with 35.9% completeness, while passive hosts are recovered with only 16.4% completeness. This difference arises because narrow nebular emission lines remain identifiable at low signal-to-noise, whereas passive galaxies rely primarily on weaker continuum and absorption features.
These results show that shallow spectroscopic follow-up does not simply reduce the total number of recovered redshifts; it systematically changes the recovered host-galaxy population. We are extending this analysis by simulating how host-dependent redshift-recovery incompleteness propagates into DES-like Hubble diagrams, and by applying this bias framework to future surveys such as TiDES and LSST-era supernova samples. This work provides a data-driven approach for identifying, quantifying, and eventually correcting spectroscopic selection effects in next-generation supernova dark-energy measurements.
| I am the presenting author | Yes |
|---|