Spatial Propagation of Weak Lensing Shear Response Corrections

TL;DR

This paper examines how shear response corrections in weak lensing measurements affect cosmic shear power spectra, comparing pixel-based and forward-modeling methods to optimize accuracy and minimize biases.

Contribution

It introduces a forward-modeling approach to propagate shear response corrections, avoiding issues related to pixel-based response matrix conditioning.

Findings

Pixel-based correction depends on a free parameter affecting noise and bias.

Forward-modeling approach avoids the need for this parameter.

Simulations demonstrate the effectiveness of the forward approach.

Abstract

In this paper we show how response function corrections to shear measurements (e.g. as required by Metacalibration) propagate into cosmic shear power spectra. We investigate a 2-sphere pixel (also known as HEALpixel') correction and a forward-modelling approach using simple Gaussian simulations. In the 2-sphere pixel-correction approach we find a free parameter that is the tolerated condition number of the local response matrices: if this is too large then this can cause an amplification of the shot noise power spectrum, if too small it can lead to a loss of area (and a possible selection bias). In contrast by forward-modelling the power spectrum this choice can be avoided. This also applies to map-based inference methods using shear-response calibrated maps.