The halo model as a versatile tool to predict intrinsic alignments

TL;DR

This paper develops a comprehensive halo-model framework to predict intrinsic alignments (IAs) of galaxies for cosmic shear surveys, explicitly incorporating central/satellite status, red/blue morphology, and luminosity and radial dependencies. It leverages MICE mocks to create realistic Stage III-like catalogs and tests two large-scale luminosity-dependence scenarios, while integrating recent observations of satellite alignment within halos. The authors find that Stage-III analyses can be robust under simpler IA models, but Stage-IV requires a redshift-dependent IA model (NLA-z) to avoid biases, with careful treatment of halo exclusion and intermediate scales. The work provides a flexible, data-informed tool to forecast IA contamination and informs priors and modeling choices for next-generation weak lensing analyses.

Abstract

Intrinsic alignments (IAs) of galaxies are an important contaminant for cosmic shear studies, but the modelling is complicated by the dependence of the signal on the source galaxy sample. In this paper, we use the halo model formalism to capture this diversity and examine its implications for Stage-III and Stage-IV cosmic shear surveys. We account for the different IA signatures at large and small scales, as well for the different contributions from central/satellite and red/blue galaxies, and we use realistic mocks to account for the characteristics of the galaxy populations as a function of redshift. We inform our model using the most recent observational findings: we include a luminosity dependence at both large and small scales and a radial dependence of the signal within the halo. We predict the impact of the total IA signal on the lensing angular power spectra, including the current uncertainties from the IA best-fits to illustrate the range of possible impact on the lensing signal: the lack of constraints for fainter galaxies is the main source of uncertainty for our predictions of the IA signal. We investigate how well effective models with limited degrees of freedom can account for the complexity of the IA signal. Although these lead to negligible biases for Stage-III surveys, we find that, for Stage-IV surveys, it is essential to at least include an additional parameter to capture the redshift dependence.

Figures (15)

• Figure 1: (a): The distribution in redshift and magnitudes of the sample we select from MICE, with the imposed cut in apparent magnitude at $r<24$. The figure illustrates the samples used in our analysis for the six redshift bins listed in Table \ref{['tab:zbin_galinfo']}. The plot shows a random selection of $1\%$ of the galaxies in the catalogue. (b): The luminosity distribution of the red central galaxy samples for the six redshift bins, colour coded as in (a).
• Figure 2: The colour-magnitude distribution of the sample. The red line shows the cut at $g-r > 0.61 - 0.0125(M_r + 19)$ we employ to isolate the red sequence.
• Figure 3: An illustration of the redshift dependence of the IA power spectrum at large scales (2-halo regime) due to the change of the fraction of red and satellite galaxies over the $z$-bins for our simulated cosmic shear survey. We plot the ratio of a 'weighted' GI power spectrum and the standard LA one. We assume a constant signal with amplitude $A=1$ (gold dashed line); incorporating the satellite fraction decreases the overall amplitude; at high redshift the fraction of satellites drops (see table \ref{['tab:zbin_galinfo']}), with a consequent increase of the signal (blue dotted line). At high redshift blue galaxies become important, suppressing the signal (red dot-dashed line). In this toy model, only red central galaxies are expected to contribute to the total signal (purple solid line).
• Figure 4: The IA signal for different values of the slope of the power law $\beta$. The observed $z-$dependence of the signal is only caused by the different galaxy samples that populate the redshift bins. Here, we only consider the luminosity of the red central population in our simulation.
• Figure 5: Overview of different estimates of the IA amplitude as a function luminosity. The best-fit relation from Joachimi2011b (blue line) for the MegaZ, SDSS L3 and L4 and SDSS LRG samples (blue downward facing triangles); Singh2015 best-fit (red line) on LOWZ (red circles) and the revised best-fit to GAMA+SDSS Main from Johnston2019 reported in the text (green line). The three individual samples used by Johnston2019 are shown as green squares (GAMA) and limegreen (SDSS Main sample). The yellow diamond indicates our best fit amplitude for the GAMA red central galaxies.