🔭 astrophysics

CoLoRe-2LPT: Lyman- $\alpha$ mock catalogues for the validation of DESI cosmological analyses

This paper introduces CoLoRe-2LPT, a new generation of fast, physically motivated Lyman- $\alpha$ mock catalogues based on second-order Lagrangian perturbation theory that significantly improve upon previous log-normal approaches by accurately reproducing key observational statistics and systematics to validate DESI's cosmological analyses.

Original authors: M. F. Ruiz-Herrera Bernal, S. Avila, A. Font-Ribera, H. K. Herrera-Alcantar, D. Alonso, A. Cuceu, L. Casas, F. Sinigaglia, J. Aguilar, S. Ahlen, O. Alves, U. Andrade, E. Armengaud, A. Bault, F. Beutle

Published 2026-07-31

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: M. F. Ruiz-Herrera Bernal, S. Avila, A. Font-Ribera, H. K. Herrera-Alcantar, D. Alonso, A. Cuceu, L. Casas, F. Sinigaglia, J. Aguilar, S. Ahlen, O. Alves, U. Andrade, E. Armengaud, A. Bault, F. Beutler, D. Bianchi, M. Bonici, A. Brodzeller, D. Brooks, A. Carnero Rosell, J. Chaves-Montero, Z. Chen, Y. Cho, T. Claybaugh, K. S. Dawson, A. de la Macorra, A. Dey, W. Elbers, S. Ferraro, J. E. Forero-Romero, G. Gambardella, S. Gontcho A Gontcho, D. Gonzalez, A. X. Gonzalez-Morales, C. Gordon, R. Gsponer, G. Gutierrez, J. Guy, C. Hahn, M. Herbold, K. Honscheid, M. Ishak, S. Jos, S. Juneau, N. V. Kamble, T. Karim, D. Kirkby, A. Kremin, M. Landriau, A. Leauthaud, Q. Li, W. Liu, K. Lodha, O. Manasoiu, M. Manera, P. Martini, A. Meisner, R. Miquel, J. Morawetz, P. Mukherjee, A. Muñoz-Gutiérrez, S. Nadathur, G. Niz, H. E. Noriega, E. Paillas, N. Palanque-Delabrouille, J. Pan, M. Pellejero Ibanez, W. J. Percival, C. Poppett, F. Prada, H. Pulido-Hernández, A. Pérez-Fernández, I. Pérez-Ràfols, C. Ravoux, J. Rohlf, G. Rossi, R. Ruggeri, L. Samushia, E. Sanchez, C. Saulder, D. Schlegel, H. Seo, J. Silber, M. Siudek, G. Tarlé, W. Turner, R. Vaisakh, M. Vargas-Magaña, B. A. Weaver, H. Yang, H. Zhang

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 the universe as a giant, expanding balloon covered in a thick, invisible fog. For decades, astronomers have tried to map the hidden patterns inside this fog to understand how the universe grows and why it's speeding up. The key to unlocking these secrets lies in looking at the "Lyman-alpha forest." Picture a distant, blazing lighthouse (a quasar) shining its light through the foggy universe toward us. As the light travels, it gets absorbed by tiny, invisible clouds of hydrogen gas, leaving a series of dark, jagged scratches on the light's spectrum. These scratches form a "forest" of absorption lines. By studying the spacing and shape of these scratches, scientists can measure the cosmic distances between galaxies and test the rules of gravity and dark energy. However, to trust these measurements, astronomers need to know exactly what the forest should look like if their theories are correct. This requires creating perfect, virtual copies of the universe to test their tools against, a task that is incredibly difficult because the gas clouds behave in complex, messy ways that are hard to simulate on a computer.

This paper introduces a new, super-smart way to build these virtual universes, specifically for the Dark Energy Spectroscopic Instrument (DESI), a massive telescope survey currently mapping the cosmos. The authors, led by M. F. Ruiz-Herrera Bernal, have developed a new generation of "mock catalogues"—essentially, highly realistic computer simulations of the Lyman-alpha forest. Instead of using older, simpler methods that treated the universe like a smooth, predictable fluid, they used a technique called "second-order Lagrangian perturbation theory" (2LPT). Think of this as upgrading from a flat, 2D map to a 3D model that can handle the gentle, wavy curves of the universe's expansion and the slight squishing of gas clouds as they clump together.

The team found that their new simulations are significantly better than the old ones. They successfully recreated the tiny, detailed clumps of gas and the broad, fuzzy peaks in the data that previous models missed. In fact, their virtual forests matched the real measurements from DESI to within 10%, capturing everything from the average amount of light getting through to the specific way the gas moves. They also improved how they simulated the quasars themselves, making sure the "lighthouses" were scattered in the sky just like the real ones. Crucially, they added realistic "contaminants" to their simulations, such as heavy metal lines and dense gas systems that can trick the analysis, ensuring the tools used to read the data won't get confused by these cosmic impostors.

The result is a powerful, fast, and physically accurate toolkit. The authors demonstrate that these new mocks can validate the complex analyses DESI uses to measure the expansion of the universe. While the simulations show a tiny, negligible bias of about 0.3% in the final cosmological numbers (likely due to the model itself rather than the simulation method), the overall agreement is strong enough to confirm that the DESI team can trust their data. This work doesn't just solve a problem; it provides the essential "ground truth" needed to ensure that the next generation of cosmic maps reveals the true nature of our accelerating universe.

Problem Statement
The Lyman-α (Lyα) forest is a critical probe for studying the large-scale structure of the universe at high redshifts ( $z > 2$ ), offering constraints on Baryon Acoustic Oscillations (BAO) and full-shape (FS) clustering. However, as the Dark Energy Spectroscopic Instrument (DESI) enters an era of unprecedented precision with its Data Release 2 (DR2), the sensitivity to modeling assumptions, astrophysical systematics, and observational effects has increased. Realistic mock catalogues are essential for validating analysis pipelines, quantifying systematic biases, and verifying parameter recovery. Previous mock generation methods for the Lyα forest relied on log-normal realizations of the density field. While effective for large-scale BAO studies, these approaches are unreliable for mildly non-linear scales ( $k \sim 0.1 - 0.5 \, h/\text{Mpc}$ ) and fail to accurately reproduce the non-linear broadening of the BAO peak or the small-scale clustering required for robust FS analyses. Furthermore, previous mocks often lacked realistic modeling of quasars (QSOs) and astrophysical contaminants like High Column Density systems (HCDs) and metal lines.

Methodology
The authors present "CoLoRe-2LPT," a new suite of fast cosmological Lyα mock catalogues designed to address the limitations of log-normal approaches. The methodology involves a multi-stage pipeline:

2LPT Density and Velocity Fields: The core innovation is the use of Second-Order Lagrangian Perturbation Theory (2LPT) to evolve the matter density field, replacing the log-normal transformation. The pipeline utilizes the CoLoRe code to generate cosmological light cones. It starts with a Gaussian random field at $z=0$ , evolves test particles using 2LPT to obtain displacements and velocities, and rescales them to generate light cones covering the DESI QSO redshift range ( $0 < z < 3.8$ ).
QSO Clustering Calibration: To populate the light cones with QSOs, the authors employ a density-dependent threshold bias model. The parameters of this model are empirically calibrated against high-resolution Abacus N-body simulations. This ensures the QSO linear bias and redshift-space distortions (RSD) match observational data. The model includes the addition of "Fingers-of-God" (FoG) effects via random velocities to reproduce the damping seen in N-body simulations.
Lyα Transmission Generation: The LyaCoLoRe module transforms the 2LPT density and velocity skewers into Lyα transmission skewers. This process involves:
- Adding small-scale power to the 2LPT density fields using a 1D power spectrum to match the forest's characteristic clustering down to $\sim 100 \, \text{kpc}/h$ .
- Converting density to optical depth using the Fluctuating Gunn-Peterson Approximation (FGPA).
- Applying RSDs to map real-space optical depth to redshift space.
- Computing transmitted flux.
- Parameters for small-scale power, optical depth, and velocity scaling are tuned to reproduce DESI DR1 measurements of the 1D power spectrum ( $P_{1D}$ ), mean flux, Lyα bias ( $b_{\text{Ly}\alpha}$ ), and the RSD parameter ( $\beta_{\text{Ly}\alpha}$ ).
Astrophysical Contaminants and Noise: To create DESI-like mocks, the pipeline incorporates:
- HCDs: Modeled by populating cells above a density threshold, with parameters tuned to match HCD clustering and incidence rates.
- Metal Lines: Added after RSDs using the quickquasars script, modeling transitions (e.g., Si II, Si III) that create spurious correlation bumps.
- Continuum and Noise: QSO continua are generated using SIMQSO, and instrumental noise is added via specsim to match the DESI DR2 footprint and exposure characteristics.

Key Contributions

First Application of 2LPT to Full Light Cone Lyα Mocks: While LPT methods have been used for galaxy surveys and box-based Lyα mocks, this work constitutes the first application of 2LPT to full light cone mocks for a Lyα forest survey.
Improved Non-Linear Modeling: The 2LPT approach naturally captures mild non-linearities and the non-linear broadening of the BAO peak, which were previously modeled ad-hoc in log-normal mocks.
Comprehensive Validation: The mocks are validated against multiple DESI statistics, including the 1D power spectrum, mean flux, Lyα and QSO biases, and the $\beta$ parameter.
Realistic Contaminant Modeling: The inclusion of HCDs and metal lines allows for the exploration of systematics induced by these astrophysical contaminants, which is crucial for FS analyses.

Results

Clustering Statistics: The new mocks reproduce Lyα statistics within 10% of the latest DESI measurements. Specifically, the Lyα bias and $\beta$ parameter, mean transmitted flux, and 1D power spectrum are well-reproduced.
QSO Clustering: The QSO linear bias is recovered to within 5% of observational values. The inclusion of 2LPT velocities and FoG effects significantly improves the recovery of redshift-space distortions compared to previous approaches.
BAO and FS Validation: When fitted with the full DESI DR2 BAO model, the mock correlations show excellent agreement with observational data. The cosmological parameters derived from the mocks deviate from unity by at most 0.003, a shift negligible compared to statistical uncertainties. The mocks successfully reproduce the anisotropic damping of the BAO peak naturally, without ad-hoc broadening.
Contaminant Validation: The inclusion of metals and HCDs results in mock correlations that qualitatively and quantitatively match the spurious bumps and distortions observed in DESI DR2 data. The metal biases agree with DESI DR2 parameters within $\sim 20\%$ .

Significance
The paper claims that this new set of mocks is a "key" tool for enhancing the modeling and validation of the DESI DR2 Lyα full-shape cosmological analysis. By providing a physically motivated and computationally efficient tool that accurately captures both large-scale and mildly non-linear clustering, the CoLoRe-2LPT mocks enable robust validation of FS and BAO analyses for current and next-generation Lyα surveys. The work demonstrates that 2LPT-based mocks can replace log-normal approaches to provide more realistic small-scale features and BAO broadening, which is essential as cosmological analyses move beyond simple BAO measurements to full-shape constraints.

More like this