Simulation Tests of PSF Modeling for Cosmic Shear with the Vera C. Rubin Observatory
📄 arXiv:2608.21688 · 📥 PDF · 2026-08-21 · astro-ph.CO
Authors: Claire-Alice Hébert [arXiv · scholar] , Erin S. Sheldon [arXiv · scholar] , Tianqing Zhang [arXiv · scholar] , Joachim Harnois-Déraps [arXiv · scholar] , Mike Jarvis [arXiv · scholar] , the LSST Dark Energy Science Collaboration [arXiv · scholar]
📄 Abstract
Exceptional control of systematic effects is required in order to achieve unbiased cosmic shear two-point correlation function measurements with the next generation of galaxy imaging surveys, such as the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST). One critical challenge is accurately modeling the point-spread function (PSF), as errors in PSF estimation can introduce spatially correlated biases in galaxy shape measurements. The LSST Science Pipelines, which will be used to process Rubin data, include an implementation of the PIFF (PSFs in the Full Field of View) package originally developed for, and demonstrated to perform well on, DES-Y3 and Y6 data. In this work we use semi-realistic image simulations, mimicking LSST observing conditions in $i$-band over 100 square degrees, to perform an end-to-end test of the Rubin PSF modeling pipeline. PSF model residuals are quantified using both second- and fourth-order moment parameters and performance is evaluated, for both LSST year 1 and year 10 depth, with a series of diagnostic tests. We find that the additive bias contribution from PSF modeling errors to the non-tomographic cosmic shear data vector is well below 30% of the cosmic shear uncertainty estimated from our analytic covariance matrix. Though our simulations exclude several known effects that may further challenge PSF modeling, these results demonstrate promising performance from PIFF on LSST-like data and provide an early benchmark for ongoing PSF validation efforts for LSST weak lensing analyses.