🔭 astrophysics

The Carousel Lens II: Cosmological Constraints with GIGA-Lens

This paper demonstrates that the Carousel Lens, a relaxed cluster-scale strong gravitational lens with multiple source planes analyzed via the GIGA-Lens pipeline, yields competitive cosmological constraints on dark matter and dark energy that rival traditional probes like the CMB and SNe Ia, with projected improvements from additional high-redshift sources and reduced systematic uncertainties.

Original authors: Felipe Urcelay, Xiaosheng Huang, William Sheu, Jackson H. O'Donnell, Tesla Jeltema, Demetrius Y. Williams, Sean Xu, Shrihan Agarwal, Greg Aldering, David Álvarez-García, Harsh Ambardekar, Tania M. Bar

Published 2026-03-02

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Felipe Urcelay, Xiaosheng Huang, William Sheu, Jackson H. O'Donnell, Tesla Jeltema, Demetrius Y. Williams, Sean Xu, Shrihan Agarwal, Greg Aldering, David Álvarez-García, Harsh Ambardekar, Tania M. Barone, Fuyan Bian, Adam S. Bolton, Aleksandar Cikota, Gerrit S. Farren, Karl Glazebrook, Taylor Hoyt, Aniket Jain, Tucker Jones, Glenn G. Kacprzak, Emerald Lin, Saul Perlmutter, David Rubin, David J. Schlegel, Ethan Silver, Christopher J. Storfer, Nao Suzuki, Jannik Truong, Mónica Úbeda, Keerthi Vasan G. C

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: Weighing the Invisible Universe

Imagine the universe is a giant, invisible ocean. We know it's there because of the waves (gravity) it creates, but we can't see the water itself. Scientists have long suspected that this ocean is made of two mysterious ingredients: Dark Matter (which acts like heavy anchors holding galaxies together) and Dark Energy (which acts like an invisible wind pushing the universe apart).

For decades, we've tried to measure how much of each ingredient is in the ocean using different tools: looking at the "baby pictures" of the universe (the Cosmic Microwave Background), watching exploding stars (Supernovae), and measuring the ripples in space (BAO). But there's a problem: our measurements don't always agree, and we still don't know exactly how much Dark Energy is pushing the universe apart.

This paper introduces a new, clever tool to solve this puzzle: The Carousel Lens.

The Tool: A Cosmic Funhouse Mirror

Think of a galaxy cluster (a massive group of galaxies) as a giant, warped funhouse mirror. When light from a distant galaxy behind it passes through this cluster, the gravity of the cluster bends the light, stretching and multiplying the image of the background galaxy. This is called Strong Gravitational Lensing.

Usually, these funhouse mirrors are messy. They are twisted, lumpy, and hard to understand. But the Carousel Lens is special. It's a "relaxed" cluster, meaning it's calm and smooth, like a well-oiled carousel. Because it's so smooth, it acts like a high-precision lens rather than a distorted funhouse mirror.

The Experiment: The "Hubble Diagram" for Gravity

The scientists used a powerful telescope (Hubble) to look at this Carousel. They found 11 different background galaxies that were being lensed. Some were close, some were far away.

Here is the magic trick:

The Geometry: Imagine you are standing on a hill (Earth). You see a tree (the lens) in the middle distance. Behind the tree, you see two birds flying at different heights (the background sources).
The Trick: The way the tree bends the light from the lower bird is different than how it bends the light from the higher bird.
The Math: By measuring exactly how much the light is bent for birds at different distances, the scientists can calculate the shape of the "ocean" (the universe) between them.

The paper used a super-computer pipeline called GIGA-Lens to build a 3D model of this lens. It's like taking a blurry photo of a funhouse mirror and using math to reconstruct exactly what the mirror looks like, pixel by pixel.

The Results: A New Way to Measure the Universe

Using this model, the team measured two key things:

$\Omega_m$ (Omega-m): How much "heavy stuff" (matter) is in the universe.
$w$ (w): How hard the "wind" (Dark Energy) is blowing.

The Current Score:
Based on the data they have right now, they found:

Matter: About 34% (close to the expected 30%).
Dark Energy: A value of -1.31 (close to the expected -1).

The Catch:
Right now, their measurement is a bit "fuzzy." It's like trying to guess the weight of a car by looking at it through a foggy window. The "fog" here is systematic uncertainty—small errors in how they modeled the lens. Currently, these errors are about 5 times bigger than the random noise in the data.

The Future: Why This Matters

The paper argues that this method is a game-changer for two reasons:

It's a Different Angle: Imagine trying to find a lost key. If you look from the North, you see a shadow. If you look from the East, you see a different shadow. The Carousel Lens looks at the universe from a completely different angle than the Cosmic Microwave Background or Supernovae. When you combine these different angles, you get a much sharper picture.
The "Zoom" Effect: The scientists realized they are currently only looking at galaxies that are "medium" distance away. They know there are even further galaxies (detected by a different instrument called MUSE) that they haven't fully used yet.
- The Prediction: If they can get high-resolution photos of these super-distant galaxies (perhaps with the James Webb Space Telescope), they can add more "birds" to their experiment.
- The Result: The paper predicts that adding these distant sources will make their measurements 80% more precise. At that point, a single galaxy cluster could tell us as much about the universe as the massive surveys of thousands of supernovae.

The Bottom Line

This paper is like a proof-of-concept for a new type of telescope. It shows that by using calm, smooth galaxy clusters as giant lenses, we can measure the expansion of the universe with incredible precision.

While the current measurements are a bit fuzzy due to modeling challenges, the potential is huge. If we can clear up the "fog" (reduce the systematic errors) and look at even more distant galaxies, the Carousel Lens could become one of the most powerful tools we have to understand the mysterious Dark Energy that is driving our universe apart.

In short: They found a calm, cosmic mirror, used a super-computer to map its curves, and proved that this mirror can measure the universe's expansion almost as well as our best existing tools—potentially even better in the future.

1. Problem Statement

The nature of dark matter and dark energy remains a central challenge in cosmology. While the $\Lambda$ CDM model is successful, tensions exist in key parameters (e.g., the Hubble constant) and uncertainties persist regarding the dark energy equation of state ( $w$ ). Traditional probes like the Cosmic Microwave Background (CMB), Type Ia Supernovae (SNe Ia), and Baryon Acoustic Oscillations (BAO) are subject to specific systematics.

Strong gravitational lensing offers a geometric probe independent of these systematics. Specifically, multi-source-plane lenses (where a single lens deflects light from background sources at different redshifts) can constrain cosmology because the deflection angle depends on angular-diameter distance ratios, which are functions of matter density ( $\Omega_m$ ) and the dark energy equation of state ( $w$ ). However, previous constraints from this technique have lagged behind other probes due to:

Galaxy-scale lenses: Limited constraining power per system and complex multi-plane effects.
Cluster-scale lenses: Often dynamically complex, requiring intricate mass models with many components, leading to significant modeling systematics.
Modeling limitations: Historically, cluster lenses were modeled using point-source positions rather than extended surface brightness, limiting precision.

2. Methodology

The authors present a new analysis of the Carousel Lens (DESI J090.9854-35.96831), a relaxed galaxy cluster at $z_l = 0.49$ .

Data:
- Imaging: Hubble Space Telescope (HST) WFC3 data in F140W and F200LP bands.
- Spectroscopy: VLT/MUSE and Gemini/GMOS data providing spectroscopic redshifts for 11 multiply imaged background sources (ranging from $z \approx 0.96$ to $z \approx 4.1$ ).
- Selected Sources: The analysis focuses on 6 HST-resolved sources (1, 3, 4, 5, 6, 7) spanning four redshift planes ( $0.96 \lesssim z \lesssim 1.6$ ) and uses source 9 to constrain a local subhalo.
Modeling Framework (GIGA-Lens):
- The study utilizes the GIGA-Lens pipeline, a GPU-accelerated, fully parametric framework using automatic differentiation (JAX) for gradient-based inference.
- Mass Model: A pixel-level model comprising four mass components (Elliptical Power-Law profiles) and external shear.
  - Main halo (centered on cluster members La/Lb).
  - Subhalos associated with galaxies Ld, Le, and a collective clump near Lf.
- Source Modeling: Extended surface-brightness modeling for sources 1–7. Complex morphologies (e.g., Source 3) are modeled using combinations of Sérsic profiles and Shapelet bases.
- Inference:
  1. MAP Stage: Multi-start gradient optimization to find the best-fit model.
  2. MCMC Stage: Hamiltonian Monte Carlo (HMC) with parallel tempering to sample the posterior distribution.
  3. Uncertainty Calibration: To address model misspecification (where the parametric model cannot perfectly reproduce high-S/N pixel data), the authors rescale the log-likelihood by a factor derived from the reduced $\chi^2_\nu$ ( $\hat{\chi}^2_\nu = 2.20$ ). This "error floor" calibration ensures statistical uncertainties are not underestimated.
Systematic Error Estimation:
- Dedicated simulations (10 mock systems) were created using more complex physics (NFW halos + dPIE subhalos) than the fiducial model.
- The deviation in deflection ratios ( $\Delta\eta$ ) between the mock truth and the fiducial model fit was used to quantify systematic uncertainties ( $\sigma_{sys}$ ).
Cosmological Inference:
- A two-step process: First, derive deflection ratios ( $\eta_i$ ) from the lens model; second, constrain cosmological parameters ( $\Omega_m, w$ ) by integrating the likelihood over a grid using the measured $\eta_i$ and their covariance.

3. Key Contributions

First Pixel-Level Cluster Cosmography: This is the first study to derive cosmological constraints from a galaxy cluster lens using extended surface-brightness modeling (pixel-level) rather than just image positions.
Relaxed System Advantage: Demonstrates that "relaxed" cluster lenses with simple inner mass distributions (dominated by a single halo + few subhalos) can yield competitive constraints compared to complex, unrelaxed clusters.
Systematic Quantification: Provides a rigorous, simulation-based quantification of systematic uncertainties arising from mass profile assumptions and substructure, showing they currently dominate statistical errors.
Forecasting: Projects the constraining power of including higher-redshift sources ( $z \sim 3-4$ ) identified in MUSE data, assuming future high-resolution imaging (e.g., JWST).

4. Results

Lens Model Performance:
- The model successfully reproduces the positions and morphologies of 6 source families (237 free parameters).
- Reduced $\chi^2_\nu = 1.38$ (within the mask).
- RMS residuals are $\sim 0.17''$ for non-blended images, comparable to or better than typical cluster lens models.
- The model predicted the redshifts of sources 6 and 7 ( $z \sim 1.6$ ) before spectroscopic confirmation, validating the reconstruction.
Cosmological Constraints ( $w$ CDM):
- Using the 6 HST sources, the authors derived:
  - $\Omega_m = 0.34^{+0.16}_{-0.13}$
  - $w = -1.31^{+0.35}_{-0.32}$
- Uncertainty Breakdown: Systematic uncertainties are roughly 5 times larger than statistical uncertainties.
- Degeneracy: The Carousel Lens constraints are nearly orthogonal to CMB constraints, making them highly complementary.
- Combined Constraints: Combining Carousel Lens with Planck 2018 CMB data yields $\Omega_m = 0.26 \pm 0.03$ and $w = -1.2 \pm 0.1$ .
Evolving Dark Energy ( $w_0w_a$ CDM):
- Constraints on $w_0$ and $w_a$ are comparable in precision to CMB alone but suffer from strong degeneracy between $w_0$ and $w_a$ due to the limited redshift leverage of current sources.
Forecasts:
- Including 4 additional high-redshift sources ( $z \sim 3-4$ ) with similar fractional uncertainties could improve constraining power by ~80%.
- With these additions, a single cluster lens could achieve precision comparable to CMB, SNe Ia, and BAO, while offering independent systematics.

5. Significance

Competitive Probe: The study demonstrates that relaxed, multi-source-plane cluster lenses are a viable and competitive probe for precision cosmology, capable of rivaling established methods when systematics are controlled.
Complementarity: The geometric nature of the probe provides constraints orthogonal to CMB and independent systematics from SNe Ia and BAO, crucial for breaking degeneracies and validating the $\Lambda$ CDM model.
Future Pathway: The results highlight the potential of JWST and future high-resolution imaging to unlock the full cosmological power of cluster lenses by resolving high-redshift sources ( $z > 3$ ) and enabling multi-band modeling.
Methodological Shift: By moving from position-based to surface-brightness-based modeling for clusters, the field can extract more information (magnification, flexion) and reduce statistical errors, paving the way for next-generation cosmological surveys.