Cosmic shear beyond 2-point statistics: Accounting for galaxy intrinsic alignment with projected tidal fields

TL;DR

This paper develops a new method to incorporate galaxy intrinsic alignments into higher-order weak lensing statistics, revealing significant biases in cosmological parameters if unaccounted for, and offers a flexible approach for future surveys.

Contribution

It introduces a novel 2D IA infusion technique for aperture mass map statistics, extending IA modeling beyond traditional two-point functions in weak lensing analyses.

Findings

Large peaks in aperture mass maps are significantly affected by IA, with deviations up to 30%.

Unmodeled IA causes substantial biases in cosmological parameters like $w_0$, $\

paper_type":"empirical"}}

Abstract

Developing analysis pipelines based on statistics beyond two-point functions is critical for extracting a maximal amount of cosmological information from current and upcoming weak lensing surveys. In this paper, we study the impact of the intrinsic alignment of galaxies (IA) on three promising probes measured from aperture mass maps -- the lensing peaks, minima and full PDF, in comparison and in combination with the shear two-point correlation functions ($\gamma$-2PCFs). Our two-dimensional IA infusion method converts the light-cone-projected mass sheets into projected tidal tensors, which are then linearly coupled to an intrinsic ellipticity component with a strength controlled by the coupling parameter $A_{\rm IA}$. We validate our method with the $\gamma$-2PCFs statistics, recovering well the analytical calculations from the linear alignment model of Bridle \& King in a full tomographic setting, and for different $A_{\rm IA}$ values. We next use our method to infuse at the galaxy catalogue level a non-linear IA model that includes the density-weighting term introduced in \citet{Blazek2015}, and compute the impact on the three aperture mass map statistics. We find that large \snr peaks are maximally affected, with deviations reaching 30\% (10\%) for a {\it Euclid}-like (KiDS-like) survey. Modelling the signal in a $w$CDM cosmology universe with $N$-body simulations, we forecast the cosmological bias caused by unmodelled IA for 100 deg$^2$ of {\it Euclid}-like data, finding very large offsets in $w_0$ (5-10$\sigma_{\rm stat}$), $\Omega_{\rm m}$ (4-6$\sigma_{\rm stat}$), and $S_8 \equiv \sigma_8\sqrt{\Omega_{\rm m}/0.3}$ ($\sim$3$\sigma_{\rm stat}$). The method presented in this paper offers a compelling avenue to account for IA in beyond-two-point weak lensing statistics, with a flexibility comparable to that of current $\gamma$-2PCFs IA analytical models.