🔭 astrophysics

A Consistent Implementation of Cluster Strong Lensing in Cosmological Simulation Light Cones

This paper presents a fully simulation-based method that generates strong-lensing images directly from particle data by remapping simulation volumes and employing multi-plane ray tracing, demonstrating that uncorrelated line-of-sight structure significantly alters image positions and critical curve morphology in galaxy cluster lensing.

Original authors: Cian Roche, Mark Vogelsberger, Michael McDonald, Isaque Dutra, Priyamvada Natarajan, Xuejian Shen, Soumya Shreeram, R. Benton Metcalf, Keren Sharon, Simon Birrer, Wonki Lee, Massimo Meneghetti

Published 2026-06-01

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Cian Roche, Mark Vogelsberger, Michael McDonald, Isaque Dutra, Priyamvada Natarajan, Xuejian Shen, Soumya Shreeram, R. Benton Metcalf, Keren Sharon, Simon Birrer, Wonki Lee, Massimo Meneghetti

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

The Big Picture: Building a Cosmic Hall of Mirrors

Imagine you are trying to understand how a giant, invisible cosmic magnifying glass (a galaxy cluster) bends light from distant stars. In the past, scientists tried to simulate this by building a "toy model": they would take a perfect, smooth mathematical shape for the lens, place a few fake stars behind it, and calculate the result.

The problem? The universe isn't a toy model. It's messy, clumpy, and full of hidden structures.

This paper introduces a new way to simulate these cosmic lenses. Instead of building a toy model, the authors take a giant, realistic video game world (a cosmological simulation called IllustrisTNG) and slice it up to create a "light cone." Think of this as peeling back layers of an onion to see everything from the center of the universe all the way out to the edge, including every galaxy, gas cloud, and dark matter clump along the way.

The Main Problem: The "Box" vs. The "Beam"

The authors faced a tricky geometry problem, like trying to fit a long, thin laser beam inside a square cardboard box.

The Box: Computer simulations usually run in big, square boxes (like a cube).
The Beam: To see a galaxy cluster lensing a distant object, you need to look a very long way into space (about 7.5 billion light-years). But you only need to look at a tiny patch of sky (a few hundred arcseconds wide).

If you tried to build a simulation box big enough to reach those distant stars and detailed enough to see the tiny stars in the background, your computer would need more power than exists on Earth. It's like trying to build a map of the entire Earth where every single blade of grass is visible; the file size would be impossible.

The Solution: The "Squeeze" Trick

To solve this, the authors used a clever trick called Box Remapping.

Imagine you have a square block of Jell-O with fruit pieces suspended inside it. You want to stretch it into a long, thin tube (like a straw) so you can look through it from one end to the other, but you don't want to lose any of the fruit or squish them together.

The authors wrote a code that mathematically "stretches" the square simulation box into a long, thin prism.

The Stretch: It stretches the box along the line of sight (the long way) so it reaches deep into the universe.
The Squeeze: It squeezes the box on the sides (the width) so it stays small enough to fit on a computer.
The Rule: Crucially, this stretch doesn't break the Jell-O. The fruit (galaxies and dark matter) stays in the same relative order. Nothing gets duplicated, and nothing gets lost. It just changes the shape of the container to match the shape of the telescope's view.

How They Made the Images

Once they had this long, thin "tube" of the universe, they did the following:

Slicing the Tube: They sliced the tube into thin layers (planes), like slicing a loaf of bread. Each slice represents a moment in time (a specific redshift) in the universe's history.
The Lens and the Source: They picked a massive galaxy cluster in the middle of the tube to be the "lens." They looked at all the galaxies behind it (the "sources").
Ray Tracing: They simulated beams of light traveling from the back of the tube, through every single slice of bread (every layer of matter), all the way to the "observer" at the front.
The Result: Because the light passed through every slice of matter—not just the main cluster—the simulation showed how the light was bent by the main cluster plus all the random, unconnected clumps of matter in the background.

What They Found: The "Ghost" Effect

The most important discovery was about the "background noise."

In the past, scientists often assumed that only the main galaxy cluster mattered. They ignored the random galaxies and dark matter floating in the space between the cluster and us.

The authors found that these "background ghosts" actually matter a lot:

Shifting Images: The random stuff in the background can push the images of distant galaxies by several arcseconds. That's like moving a dot on a screen by a noticeable amount.
Warping the Magnifying Glass: The "critical curve" (the ring where the magnification is infinite) isn't a perfect circle. The background stuff makes it wobble, change size, and sometimes even break into smaller pieces.
The Numbers: They found that ignoring this background stuff can change the size of the main magnifying ring by about 6% on average, but in some cases, it can change it by as much as 50%.

The Conclusion

This paper proves that we can now generate realistic images of strong gravitational lensing directly from a single, consistent computer simulation. We don't need to mix and match different models anymore.

The key takeaway is that the universe is connected. To understand how a galaxy cluster bends light, you can't just look at the cluster; you have to look at the entire "light cone" of the universe behind it. If you ignore the messy, unconnected stuff in the background, your predictions for where the images will appear will be off, which could mess up our measurements of the universe's expansion and the nature of dark matter.

In short: They built a better, more realistic "cosmic camera" that shows us that the background noise is actually part of the signal.

Technical Summary: A Consistent Implementation of Cluster Strong Lensing in Cosmological Simulation Light Cones

Problem Statement
While galaxy cluster strong gravitational lensing is a cornerstone of precision cosmology, robust theoretical predictions have historically lagged behind high-quality observations. This gap stems from two primary issues: the geometric mismatch between standard cubic simulation boxes and the elongated geometry required for light cones, and the fundamental tension between the large volume needed to host massive clusters and the high resolution required to resolve background source galaxies and substructure.

Current forecasting methods often rely on hybrid approaches, extracting individual lenses from simulations and combining them with analytic or observed source populations. These methods frequently fail to preserve correlations across redshift or omit correlated and uncorrelated line-of-sight (LoS) structure. Consequently, there is a lack of simulated strong lensing images generated consistently from simulation outputs, where the primary lens, sources, and all intervening matter are drawn from the same cosmological volume in a manner faithful to the standard cold dark matter paradigm.

Methodology
The authors present a fully simulation-based procedure to generate strong-lensing images directly from particle data, utilizing the IllustrisTNG cosmological hydrodynamical simulations. The methodology consists of three main components:

Volume Remapping: To address the geometric constraints, the authors employ a box remapping technique (using the BoxRemap code) to transform the cubic simulation volume into a right square prism. This geometry features two short axes (representing the plane of the sky) and one long axis (representing the line of sight). This transformation preserves total volume and local structure without introducing repeated structures within a single light cone, provided the field of view is chosen appropriately relative to the box size. The remapping is applied consistently to every snapshot of the simulation.
Plane Extraction and Smoothing: From the remapped boxes, the authors extract mass and light information to create a series of lens and source planes corresponding to simulation snapshots.
- Mass Planes: Particles are divided into radial shells based on snapshot redshifts. Mass maps are generated by smoothing dark matter, star, and gas particles using a nearest-neighbor kernel (32 neighbors, Wendland-C2 kernel). To account for diffuse mass in sparse regions and avoid shot noise, the total mass of the shell is added as a uniform sheet, and the mean matter density of the universe is subtracted to create a mass contrast map.
- Source Planes: Stellar particles are smoothed into flux maps for specific observational filters (e.g., JWST f200w). The flux is estimated based on stellar age, initial mass, and metallicity, redshifted to the source plane, and converted to apparent magnitude. Subhalos with insufficient particle counts are replaced with analytic S'ersic profiles to ensure completeness.
Multi-Plane Ray Tracing: The extracted planes are used in a multi-plane ray tracing algorithm (using the GLAMER code). Light from source planes at redshift $z_s$ is traced backward through all lens planes at redshifts $z_l < z_s$ . The resulting lensed images are stacked to produce a final theoretical prediction. The pipeline requires no manual injection of sources or analytical profiles; the cosmology and galaxy formation model of the simulation serve as the prior.

Key Contributions and Results
Using the TNG300-1 simulation, the authors generated light cones centered on massive galaxy groups ( $M_{200m} \ge 5 \times 10^{14} M_\odot$ ) at a lens redshift of $z_G \approx 0.38$ , extending to a source redshift of $z_{max} \approx 4.18$ .

Image Generation: The pipeline successfully produces strong lensing features, including multiple image configurations, arcs, and critical curves, directly from the simulation data without ad-hoc source injection. The resulting images show typical features such as tangential arcs and complex critical curve morphologies driven by cluster mergers and substructure.
Impact of Line-of-Sight Structure: The study quantifies the impact of uncorrelated LoS structure by comparing full light cone ray tracing against ray tracing through the primary lens plane only.
- Image Positions: Uncorrelated LoS structure can shift the relative positions of lensed images by several arcseconds (up to $\sim 10''$ in extreme cases), significantly exceeding typical strong lens mass modeling residuals ( $\lesssim 1''$ ).
- Critical Curve Morphology: The inclusion of LoS material introduces a $\sim 6\%$ scatter in the area of a cluster's primary critical curve. In some cases, the area of the primary critical curve was enhanced by 30% or decreased by 50%.
- Total Critical Area: Within $100''$ of the cluster potential minimum, the total critical area at $z_s = 4$ changes by $16^{+20}_{-14}\%$ when LoS structure is included. This enhancement arises from both the creation of new critical curves by LoS objects and the modification of existing curves.

Significance and Limitations
The paper claims to provide the first demonstration of cluster strong lensing images generated self-consistently from a cosmological hydrodynamical simulation light cone. This approach offers a more automated and "hands-off" method for producing large catalogs of mock observations, which can be used to study systematics in lens modeling, biases in mass inference, and the galaxy-galaxy strong lensing (GGSL) cross-section tension.

The authors note several limitations:

Large-Scale Power: The method does not introduce large-scale power modes absent from the initial simulation box, potentially biasing the sampling of the large-scale power spectrum.
Resolution: The method is limited by the simulation's mass resolution, which struggles to simultaneously resolve massive clusters and low-mass substructure/sources.
Diffuse Mass: Diffuse mass is approximated as a uniform sheet rather than preserving its specific structure, though this is necessary to avoid shot noise.
Computational Cost: Generating a single light cone requires processing every particle in the simulation box at every snapshot, which is computationally expensive and generates significant data storage requirements.
External Shear: The current implementation does not include external shear from large-scale structure outside the light cone, though the expected effect is estimated to be small ( $\gamma_{ext} \sim 0.02$ ).

The authors conclude that incorporating full light-cone information is crucial for theoretical predictions and that carefully accounting for LoS structure is necessary when interpreting observations, as uncorrelated structure can significantly alter image configurations and critical curve properties.