The impact of weak lensing on Type Ia supernovae luminosity distances

TL;DR

This paper investigates how weak gravitational lensing affects Type Ia supernova luminosity measurements and cosmological parameter estimation, proposing a method to de-lens supernova magnitudes and assessing its impact on cosmological inferences.

Contribution

It introduces a test for line-of-sight matter density bias in SN Ia observations and modifies the Tripp estimator to account for lensing, reducing scatter in the Hubble diagram.

Findings

Moderate over-density bias detected in two redshift bins.

De-lensing reduces scatter in supernova luminosity residuals.

Cosmological parameters are only slightly affected by lensing correction.

Abstract

When Type Ia supernovae are used to infer cosmological parameters, their luminosities are compared to those from a homogeneous cosmology. In this note we propose a test to examine to what degree SN Ia have been observed on lines of sight where the average matter density is \textit{not} representative of the homogeneous background. We apply our test to the Pantheon SN Ia compilation, and find two redshift bins which indicate a moderate bias to over-density at $\sim 2\sigma$. We modify the Tripp estimator to explicitly de-lens SN Ia magnitudes, and show that this reduces scatter of Hubble diagram residuals. Using our revised Tripp estimator, the effect on cosmological parameters from Pantheon in $\Lambda$CDM is however small with a change in mean value from $\Omega_{\rm m} = 0.317 \pm 0.027$ (baseline) to $\Omega_{\rm m} = 0.312 \pm 0.025$ (de-lensed). For the Flat $w$CDM case it is $\Omega_{\rm m} = 0.332 \pm 0.049$ and $w = -1.16 \pm 0.16$ (baseline) versus $\Omega_{\rm m} = 0.316 \pm 0.048$ and $w = -1.12 \pm 0.15$ (de-lensed). We note that the effect of lensing on cosmological parameters may be larger for future high-z surveys.