Controlling weak-lensing shear biases from undetected galaxies in the era of Stage IV Surveys 2023 Controlling weak-lensing shear biases from undetected galaxies in the era of Stage IV Surveys – B
L. M. Voigt · 2025
Read Controlling weak-lensing shear biases from undetected galaxies in the era of Stage IV Surveys 2023 Controlling weak-lensing shear biases from undetected galaxies in the era of Stage IV Surveys – B by L. M. Voigt (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-00666.txt.
Introduction
L Voigt wrote a 2025 cosmology paper on weak lensing, the slight bending of light from far galaxies by matter in between. Measuring that bend, called shear, helps map dark matter and test models of the universe. New Stage IV surveys will measure shear with great care and detail.
Voigt focuses on a sneaky source of error. Galaxies too faint to be detected still add their light and can bias the shear values. Using a machine learning method built to avoid noise bias, the author measures that effect for a survey like Euclid. The results help planners keep faint galaxies from skewing their maps.
Catalog record
| CREATOR | L. M. Voigt |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 107 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-00666
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Controlling weak-lensing shear biases from undetected galaxies in the era of Stage IV Surveys 2023 Controlling weak-lensing shear biases from undetected galaxies in the era of Stage IV Surveys – B L. M. Voigt Abstract Gravitational lensing of background galaxies by intervening matter is a powerful probe of the cosmological model. In the era of Stage IV surveys, contamination from galaxies below the detection threshold has emerged as a significant source of bias. Adopting a noise-bias-free machine-learning method to estimate shear, we quantify the impact of faint galaxies for a Euclid-like survey. In our baseline simulations, faint blends induce a multiplicative shear bias of -0.008, well above Euclid’s requirement. Similar to previous studies, we find that calibration simulations must include neighbouring galaxies to AB apparent magnitudes as faint as 27.0 (+2.1, -0.9) and within approximately 1.0 (+0.2, -0.2) arcsec of each bright sample galaxy (BSG; the galaxy for which shear is measured). By varying faint galaxy properties, we identify which ones significantly affect shear biases and quantify how well they must be constrained. Crucially, we find that biases not only depend on the mean projected faint-galaxy density and apparent-magnitude distribution across the sample, but also on how these quantities vary with the observed brightness of the BSG. Furthermore, biases are sensitive to radial and tangential alignments and positional anisotropy of faint galaxies relative to BSGs. By contrast, shear coherence between BSGs and faint galaxies, parallel orientation alignments, and variations in the faint galaxy size–magnitude relation have negligible impact within the parameter ranges explored. Our results guide calibration simulations and highlight the critical role of deep observations in measuring the properties of faint galaxies. 1 Introduction Weak gravitational lensing provides a powerful probe of cosmic structure, making it one of the main science drivers for large-scale cosmological surveys. Stage IV missions such as Euclid 1 1 1 https://www.euclid-ec.org/ ( Laureijs et al. 2011 ; Amendola et al. 2013 ) , the Vera C. Rubin Observatory’s Legacy Survey of Space and Time 2 2 2 https://rubinobservatory.org/ ( LSST Dark Energy Science Collaboration 2012 , LSST;) , and NASA’s Nancy Grace Roman Space Telescope 3 3 3 https://roman.gsfc.nasa.gov/ ( Spergel et al. 2015 ) aim to deliver sub-percent level constraints on cosmological parameters, including the dark energy equation-of-state parameter w_{0} . Achieving this level of precision requires stringent control of systematic biases in weak lensing measurements. Sources of systematic bias include galaxy shape measurement ( Heymans et al. 2006 ; Massey et al. 2007 ; Kitching et al. 2012 ; Mandelbaum et al. 2015 ) , photometric redshifts ( Hildebrandt, H. et al. 2010 ) , intrinsic alignments ( Hirata & Seljak 2004 ; Joachimi & Bridle 2010 ; Heymans et al. 2013 , e.g.) , and the modelling of non-linear structure and baryonic effects (see also Mandelbaum 2018 , for a review) . Shape measurement systematics include model-fitting bias ( Voigt & Bridle 2010 ) , noise bias ( Kacprzak et al. 2012 ; Refregier et al. 2012 ) , and modelling of the point spread function ( Paulin-Henriksson et al. 2008 , PSF;) , including wavelength-dependent effects ( Cypriano et al. 2010 ; Voigt et al. 2012 ) . A bias that has emerged more recently as a significant concern is blending – flux contamination from neighbouring or overlapping galaxies that affects both the detection and shape measurement of sources ( Samuroff et al. 2018 , e.g.) . These galaxies may be physically associated with the source (e.g., within groups or clusters) or unrelated galaxies projected at small angular separation. While some bright blends—i.e., those above the detection threshold—can be mitigated using deblending algorithms ( Arcelin et al. 2020 ; Zhang et al. 2024 , e.g.) or catalogue-level flagging based on SExtractor ’s output parameters ( Bertin 2011 ; Zuntz et al. 2018 ) , galaxies below the detection threshold cannot be directly removed yet still contribute flux to the images of galaxies used for shear estimation. We refer to the galaxies used for shear estimation as bright sample galaxies (BSGs), and their undetected neighbours as faint galaxies. These faint blends, despite being undetected, can introduce significant shear biases if their presence is ignored i.e., if the bright galaxy used for shear estimation is assumed to be isolated. Stage III surveys, for example the Dark Energy Survey (DES) 4 4 4 https://www.darkenergysurvey.org/ Year 3 analysis, have incorporated deep image simulations to model these effects, extending the simulated galaxy population to magnitudes as faint as m_{\rm AB}\sim 27.5 ( MacCrann et al. 2022 ) . Although some shear measurement techniques aim to reduce dependence on simulations, such as metacalibration and metadetection ( Huff & Mandelbaum 2017 ; Sheldon & Huff 2017 ; Sheldon et al. 2023 ) , simulations nevertheless remain a key component of weak lensing analyses—particularly for those using model-fitting methods such as lens fit ( Miller et al. 2007 ; Li et al. 2023 ) . In such cases, assumptions about the faint galaxy population can affect shear calibration and shift cosmological constraints. However, the extent to which the properties of the faint galaxy population in these simulations must match those in the data still requires further investigation. In this paper, we carry out a systematic study of the biases arising from faint blends, using Euclid as the reference survey. We quantify the sensitivity of the shear bias to several faint galaxy properties, such as the limiting magnitude to which galaxies must be simulated and the maximum separation from the BSG at which faint blends still contribute significantly—parameters that have been explored in previous studies using separate galaxy simulation and shear measurement methods ( Hoekstra et al. 2017 ; Euclid Collaboration: Martinet et al. 2019 ; Hoekstra, Henk et al. 2021 , hereafter H17, M19 and H21, respectively) . We extend this work to also include additional properties such as the angular distribution and orientation of faint galaxies relative to the BSG, the shear applied to faint galaxies, the slope of the faint galaxy apparent magnitude distribution, and the form of the size–magnitude relation. Crucially, we also investigate the impact of correlations between the properties of faint blends and those of the BSG—specifically, dependencies of the local faint galaxy density and magnitude distribution slope on BSG brightness. To our knowledge, this is the first time such correlations have been explicitly quantified in the context of shear calibration. These effects are particularly relevant for Stage IV surveys, where even sub-percent level biases can impact cosmological inferences. We simulate galaxy images with a fixed PSF using simplified models, with the aim of isolating the impact of the faint galaxy population on shear bias. To measure shear, we use the convolutional neural network (CNN) method developed in Voigt 2024 , which employs a committee of shallow CNNs trained to recover unbiased shear estimates in the presence of noise. This supervised learning approach relies on a training set of simulated galaxies with known shears, from which the model learns to predict shear for new data. Model-fitting and PSF-related biases are avoided by using consistent galaxy and PSF models in both the training and test simulations. To isolate biases arising from the faint galaxy population, we deliberately exclude faint galaxies from the training simulations