🔭 astrophysics

Forward analytical model for the optical selection bias on galaxy cluster lensing profiles

This paper presents a fully predictive forward analytical model that quantifies how optical selection biases, driven by line-of-sight projection effects, distort galaxy cluster lensing profiles and density statistics, successfully validating the framework against simulated DES Y3-like data to link selection biases directly to underlying cosmology.

Original authors: M. Costanzi, H. Y. Wu, J. H. Esteves, S. Grandis, C. To, M. Aguena

Published 2026-04-08

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: M. Costanzi, H. Y. Wu, J. H. Esteves, S. Grandis, C. To, M. Aguena

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to weigh a single, specific apple in a grocery store. You want to know exactly how heavy it is so you can use that weight to understand the health of the entire orchard.

In the world of astronomy, galaxy clusters are those apples. They are massive groups of hundreds or thousands of galaxies held together by gravity. Astronomers use them to measure the universe's expansion and the nature of dark energy. To do this, they need to know the cluster's true mass.

However, there's a problem: Optical telescopes (like the one used in the Dark Energy Survey) don't see mass directly. They see light. So, to guess a cluster's mass, astronomers count how many galaxies they see in a pile. This count is called "richness."

The Problem: The "Crowded Room" Effect

Here is the catch: Our view of the universe is a 3D space squashed into a 2D photo. When we look at a galaxy cluster, we are looking through a long, narrow tube (our line of sight).

Sometimes, other galaxies that don't belong to the cluster happen to line up perfectly behind or in front of it. It's like looking through a window at a party, but there's a second party happening in the street outside, and the people from the street are visible through the glass.

The Illusion: The telescope sees the party guests plus the people on the street. It thinks, "Wow, this is a huge, rich cluster!"
The Reality: The cluster is actually smaller, but it just got a "richness boost" from the background noise.
The Consequence: Because the cluster looks richer than it is, astronomers might think it's more massive than it really is. This creates a selection bias. If we don't fix this, our calculations for the entire universe will be slightly off.

The Solution: A "Forward Model"

The authors of this paper (Costanzi et al.) built a new mathematical tool—a forward analytical model—to fix this.

Think of their model as a smart detective's notebook. Instead of just guessing how much the "street people" are messing up the count, the detective calculates exactly how much they should mess it up based on:

How crowded the street is (the density of the universe).
How far away the street is (redshift).
How big the window is (the telescope's resolution).

They created a formula that says: "If a cluster looks this rich, but we know the universe is structured this way, then the 'true' richness is likely X, and the 'fake' richness from the line-of-sight is Y."

The Key Analogy: The "Echo" in a Canyon

Imagine you are shouting in a canyon. You want to know how loud your voice is. But, because of the canyon walls, your shout bounces back as an echo.

The True Voice: The actual galaxy cluster.
The Echo: The extra galaxies projected along the line of sight.
The Bias: If you just listen to the total sound (Voice + Echo), you think you are shouting louder than you actually are.

The authors' model is like a noise-canceling algorithm. It listens to the total sound, analyzes the shape of the canyon (the structure of the universe), and mathematically subtracts the echo to tell you exactly how loud your original voice was.

What Did They Find?

They tested their model using a "fake universe" (a computer simulation) that mimics the real data from the Dark Energy Survey.

It Works: Their model successfully predicted exactly how much the "echo" (projection effects) was inflating the cluster counts.
It's Scale-Dependent: The bias isn't the same everywhere. It's strongest right next to the cluster (where the echo is loudest) and fades out as you move further away, but it still lingers for a long distance.
It's Predictive: They showed that if you know the "richness boost" (how much extra stuff is in the line of sight), you can predict exactly how the cluster's density profile (its shape) is distorted.

Why Does This Matter?

In the past, astronomers had to guess how much this "crowded room" effect messed up their data, often leaving a big margin of error.

This paper provides a precise recipe to remove that error. By using their model, cosmologists can:

Weigh galaxy clusters more accurately.
Get better measurements of Dark Energy and Dark Matter.
Stop worrying that their "apples" are actually just "apples plus a pile of oranges" from the background.

In short, they built a better pair of glasses for the universe, allowing us to see the true shape of the cosmos, rather than a distorted reflection of it.

1. Problem Statement

Galaxy clusters detected via optical surveys (e.g., DES, Euclid) are subject to significant selection biases caused by projection effects. Because optical cluster finders identify clusters as overdensities of red-sequence galaxies along the line of sight (LOS), systems aligned with the LOS or containing foreground/background structures are preferentially selected.

The Core Issue: These projection effects artificially boost the observed richness ( $\lambda_{ob}$ ) of a cluster and contaminate its projected density profile ( $\Sigma$ ).
Consequence: If unmodeled, this leads to a systematic overestimation of the cluster mass (by 10–20% on scales relevant for weak lensing calibration) and introduces biases in inferred cosmological parameters.
Gap: Previous approaches relied on simulation-based calibrations or assumed analytical templates with free parameters. There was a lack of a fully predictive, parameter-free analytical model that links selection bias directly to the underlying halo occupation distribution (HOD), cosmology, and survey specifications.

2. Methodology

The authors propose a forward analytical framework to quantify the impact of LOS projection on cluster density profiles. The model is built upon the premise that the same structures boosting the observed richness also enhance the projected density profile.

A. Theoretical Framework

Richness Decomposition: The observed richness is modeled as $\lambda_{ob} = \lambda_{tr} + \Delta_{prj}$ , where $\lambda_{tr}$ is the intrinsic richness and $\Delta_{prj}$ is the contamination from projected halos.
Scale-Dependent Bias ( $b_{\lambda_{ob}}$ ): The authors introduce a scale-dependent clustering bias for optically selected clusters. They model the bias as a function of the projected angular separation ( $\theta$ $θ$ ) from the cluster center using a sigmoidal function:
$b_{\lambda_{ob}}(\theta) = b_0 + \frac{b_\infty - b_0}{1 + e^{-k(\theta - \theta_0)}}$
- $b_0$ : Small-scale bias (enhanced by projection).
- $b_\infty$ : Large-scale bias (asymptotic value).
- The transition is governed by the cluster's angular size $\theta_\lambda$ .
Two-Halo Term Derivation: The projected density profile is split into a one-halo term (intrinsic) and a two-halo term (correlated structures). The two-halo term is calculated by integrating the contributions of off-axis halos along the LOS, weighted by the derived bias $b_{\lambda_{ob}}$ and a projection kernel that accounts for redshift separation and overlapping area.
Parameter-Free Prediction: The model requires no free fitting parameters for the bias shape. Instead, the large-scale bias ( $b_\infty$ ) is derived linearly from the richness excess ( $\delta_{prj}$ ) relative to a random mass-selected sample.

B. Case Study and Validation

Survey Simulation: The model was tested against a synthetic DES Year 3 (Y3) cluster catalog generated using the redMaPPer algorithm.
Data Source: The mock catalog was built from the DeRose et al. (2019) N-body simulation (flat $\Lambda$ CDM cosmology).
Implementation:
- Intrinsic richness was assigned based on mass using a Halo Occupation Distribution (HOD).
- Observed richness ( $\lambda_{ob}$ ) was calculated by summing contributions from projected halos, accounting for redshift weights and geometric overlap.
- Projected density profiles were generated by sampling dark matter particles in cylinders around halos.

3. Key Contributions

Analytical Derivation of Bias: The paper provides the first analytical derivation of the optical cluster bias ( $b_{\lambda_{ob}}$ ) as a function of richness contamination ( $\Delta_{prj}$ ) and angular scale, without relying on simulation-based tuning for every survey configuration.
Sigmoidal Parametrization: Introduction of a sigmoid function to describe the transition from small-scale (highly biased due to projection) to large-scale bias, motivated by the correlation function ratios of richness-selected vs. random-selected samples.
Linking Richness to Environment: The model explicitly demonstrates that clusters with high richness boosts ( $\lambda_{ob} \gg \lambda_{tr}$ ) reside in significantly denser environments, leading to a higher two-halo term in their lensing profiles.
Practical Fitting Function: To facilitate use in cosmological inference pipelines (which may find the full analytical integration computationally expensive), the authors provide a double power-law fitting function with smooth transition to approximate the selection bias on the density profile.

4. Results

The model was validated against the synthetic DES Y3 mock catalog across various richness ( $\lambda_{ob}$ ) and redshift ( $z$ ) bins:

Bias Reproduction: The analytical model successfully reproduces the scale dependence and amplitude of the selection bias observed in the mock data.
Two-Halo Component: The model correctly predicts that the two-halo component of the density profile is enhanced for richness-selected clusters compared to random mass-selected samples. This enhancement correlates with the magnitude of $\Delta_{prj}$ .
Profile Distortion: The total density profile of optically selected clusters shows a characteristic bell-shaped bias relative to random samples, peaking at $\sim 1 h^{-1}$ Mpc (the transition scale between one-halo and two-halo terms).
Evolution: The model captures the mild evolution of the bias with redshift (driven by photo- $z$ uncertainty) and richness (richer clusters are more massive and thus have denser environments).
Accuracy: The model predictions match the mock data measurements within $1\sigma$ statistical uncertainties up to scales of $R \sim 10 h^{-1}$ Mpc.

5. Significance and Implications

Cosmological Precision: This framework provides a consistent methodology to correct for systematic biases in weak lensing mass calibration. By linking selection biases directly to cosmology and survey specs, it reduces the reliance on empirical calibrations that may not generalize across different surveys (e.g., LSST, Euclid).
Scalability: The analytical nature of the model allows it to be adapted to different cluster finders (e.g., WaZP) and mass proxies (e.g., stellar mass) by adjusting the projection kernel and richness distribution.
Future Applications: The derived fitting function enables efficient marginalization over projection systematics in cosmological parameter inference pipelines, a critical step for next-generation surveys where statistical errors will be subdominant to systematic uncertainties.
Limitations: The current analysis neglects cluster miscentering, halo orientation, and triaxiality. While these are sub-dominant systematics in the current context, they may require inclusion for future precision clustering analyses.

In summary, this work establishes a robust, physics-based analytical tool to quantify and correct for the optical selection bias in galaxy cluster lensing profiles, directly addressing a major systematic hurdle in modern cluster cosmology.