Description
Modeling astrophysical systematics for field-level weak lensing and galaxy clustering analysis
I will present a novel field-level pipeline for the joint analysis of weak lensing and galaxy clustering data. Our method employs a "2.5-dimensional" field-level modeling approach, where the density field is represented in slabs of approximately 100 Mpc thickness. This strategy preserves the efficiency of 2D modeling while enabling flexible treatment of key astrophysical systematics.
I will focus on modeling two critical systematics: galaxy bias and intrinsic alignment (IA). We show that field-level modeling of IA leads to more robust cosmological inference, achieving up to 5x tighter constraints on non-linear IA parameters. Additionally, I will discuss recent progress on modeling field-level galaxy bias, where we find that capturing small-scale clustering requires moving beyond the traditional Poisson likelihood to a non-Poissonian framework.
