Limitations of model fitting methods for lensing shear estimation

TL;DR

This paper identifies fundamental limitations of model-fitting methods for measuring gravitational lensing shear, showing biases arise from oversimplified galaxy models and suggesting more flexible or prior-informed approaches could improve accuracy.

Contribution

It demonstrates the fundamental bias limits of elliptical isophote model-fitting techniques and discusses how incorporating complex galaxy structures or priors can mitigate these biases.

Findings

Biases are significant when using simple elliptical models for complex galaxies.

Allowing flexible profiles reduces biases in noise-free simulations.

Biases are most pronounced for spiral-like galaxy morphologies.

Abstract

Gravitational lensing shear has the potential to be the most powerful tool for constraining the nature of dark energy. However, accurate measurement of galaxy shear is crucial and has been shown to be non-trivial by the Shear TEsting Programme. Here we demonstrate a fundamental limit to the accuracy achievable by model-fitting techniques, if oversimplistic models are used. We show that even if galaxies have elliptical isophotes, model-fitting methods which assume elliptical isophotes can have significant biases if they use the wrong profile. We use noise-free simulations to show that on allowing sufficient flexibility in the profile the biases can be made negligible. This is no longer the case if elliptical isophote models are used to fit galaxies made up of a bulge plus a disk, if these two components have different ellipticities. The limiting accuracy is dependent on the galaxy shape but we find the most significant biases for simple spiral-like galaxies. The implications for a given cosmic shear survey will depend on the actual distribution of galaxy morphologies in the universe, taking into account the survey selection function and the point spread function. However our results suggest that the impact on cosmic shear results from current and near future surveys may be negligible. Meanwhile, these results should encourage the development of existing approaches which are less sensitive to morphology, as well as methods which use priors on galaxy shapes learnt from deep surveys.