Probing Systematic Bias in Low-Redshift Type Ia Supernova Measurements by Cross Analyzing Surface Brightness and Hubble Residuals

TL;DR

This study investigates potential systematic biases in low-redshift Type Ia supernova measurements by analyzing correlations between surface brightness and Hubble residuals, finding no significant bias but highlighting its importance for future dark energy studies.

Contribution

It introduces a method to test for biases in supernova measurements by comparing surface brightness and residuals, providing a framework for future bias assessments.

Findings

No significant difference in Hubble residuals between bright and faint surface brightness halves.

The observed difference in residuals is consistent with zero at the 2 sigma level.

Future larger datasets could reveal systematic biases impacting dark energy measurements.

Abstract

For low-redshift (z < 0.1) Type Ia supernovae (SN Ia) samples used in several cosmological analyses over the past decade, we probe for systematic bias by looking for correlations between surface brightness (SB) measurements and Hubble residuals (HR). For 292 SNe Ia, we measure SB at the location of the SN Ia from publicly available Pan-STARRS (PS1) images. The Hubble residuals are from two recent measurements with low-z SNe Ia that overlap the PS1 footprint: 1) the DES 3-year cosmology analysis, with 120 overlapping low-z SNe Ia from the Harvard-Smithsonian Center for Astrophysics surveys and Carnegie Supernova Project, and 2) the PS1 single-telescope analysis, with 172 overlapping low-z SNe Ia from the Foundation Supernova Survey. This study is motivated by previous reports of anomalous inefficiencies and flux scatter for transients on bright galaxies. We compare HR distributions of the bright and faint halves of the SB distribution: the mean HR values differ by Delta HR = 0.031 +- 0.018, consistent with no difference at the 2 sigma level. We also perform a Kolmogorov-Smirnov (KS) test for the bright and faint half HR distributions, and conclude that the two distributions are statistically consistent with a KS p-value of 0.07. However, if future studies with larger datasets find Delta HR ~ 0.03 with high significance, this difference would be a leading systematic uncertainty in measurements of the dark energy equation of state, w.