Assessing the large-scale angular clustering of UNIONS Lyman Break Galaxies via cross-correlations

arXiv:2607.09943 · astro-ph.CO · Submitted 2026-07-10 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: Next we'll be talking about the paper "Assessing the large-scale angular clustering of UNIONS Lyman Break Galaxies via cross-correlations".

Jocelyn: The paper was written by the authors from.

Vera: Stay tuned as we take you through the paper and discuss its implications.

The Summary: Vera: So, taking a look at the abstract and summary of "Assessing the large-Scale Angular Clustering of UNIONS Lyman Break Galaxies via Cross-Correlations," it’s clear that they are showing a fundamental challenge in using these LBGs for auto-clustering studies.

Jocelyn: The main finding is that the way we observe these galaxies—the imaging systematics like depth and seeing—is strongly affecting the density map, making a reliable auto-angular power spectrum really difficult to obtain.

Subrahmanyan: That limitation is significant because if we can’t accurately measure how they cluster amongst themselves, our ability to test models of large-scale structure becomes much harder.

Vera: The authors are essentially saying that relying on the LBGs' own clustering is a recipe for trouble because of those image artifacts, which are quite pervasive across the entire UNIONS footprint.

Jocelyn: And instead of giving up, they’ are pivoting to cross-correlation measurements with external data sources like the Planck CMB lensing convergence map and quasars from DESI DR1 and Quaia.

Subrahmanyan: This is a huge technical shift that allows them to bypass those localized imaging problems while still accessing the cosmological information encoded in these high-redshift tracers.

Vera: It’s an elegant solution, but it opens up a whole new set of questions about how reliable the cross-correlation signal will be compared to traditional methods.

Jocelyn: We need to see how this approach performs against actual data versus how it behaves in mock simulations before we can judge the results.

Improvements and Methodology: Vera: Building on that shift, when we look at the methodology described in "Assessing the large-Scale Angular Clustering of UNIONS Lyman Break Galaxies via Cross-Correlations," they are proposing some specific techniques to manage those systematics.

Jocelyn: The paper details how they apply mitigation strategies, like using linear regression and other techniques to try and clean up the observed LBG density maps.

Subrahmanyan: I'm particularly interested in the 'namaster' code mentioned, which seems designed to remove systematic power on large scales—it's a sophisticated way of correcting instrumental biases.

Vera: It’s fascinating how they are trying to model and subtract these spatial variations, especially since the density of faint galaxies is so sensitive to subtle changes in survey depth.

Jocelyn: The paper highlights that while these correction methods help, they are essentially showing us where the limitations lie, demonstrating that the auto-spectrum is largely unsuitable for reliable cosmological inference.

Subrahmanyan: It’s a necessary honesty in this work; the theoretical models need clean data to perform their predictions, and this paper is proving that simply using the raw data doesn' problematic.

Vera: It’s clear they are being very conservative about what makes a valid measurement, which is good news for us as an observational astronomers who rely on robust results.

Jocelyn: This approach really emphasizes the cross-correlation signal as a powerful and reliable alternative, which is something that we should be looking at.

The Cross-Correlation Signal: Vera: Now, moving into the actual results of "Assessing the large-Scale Angular Clustering of UNIONS Lyman Break Galaxies via Cross-Correlations," the authors present a clear picture regarding the cross-correlation signal between UNIONS LBGs and external datasets.

Jocelyn: They show that this signal is measurable and consistent with theoretical predictions, which is a huge step toward achieving a robust measurement for these high-redshift tracers.

Subrahmanyan: This agreement with theory suggests that the LBG population is indeed tracing the underlying matter distribution in a physically coherent manner, which provides strong validation for our models of structure.

Vera: Even though there are still some residual systematics at large scales, the data shows that the cross-correlation signal is far more robust than we initially feared.

Jocelyn: The cross-correlation with Planck CMB lensing and QSO samples is demonstrating a positive correlation across a broad range of angular scales, which is exciting for any survey researcher.

Subrahmanyan: It’s also important to note that this result validates the entire UNIONS LBG sample as a reliable cosmological probe at redshift z ∼ two point five, which is exactly what we were hoping to find out.

Vera: This finding of detecting the signal is a significant contribution, and it sets up a clear direction for future work in terms of how these LBGs can be used.

Conclusion and Wrap-up: Jocelyn: So, as we wrap up our discussion on "Assessing the large-Scale Angular Clustering of UNIONS Lyman Break Galaxies via Cross-Correlations," it's clear that this paper has provided a very detailed look at the challenges and successes of using these high-redshift tracers.

Vera: It highlights that while the LBGs are promising, their own clustering is messy, but cross-correlating them makes the data much more reliable for cosmological interpretation.

Subrahmanyan: This work confirms that by utilizing cross-correlation techniques, we can extract valuable information about local primordial non-Gaussianity and structure growth in a way that overcomes significant imaging systematics.

Jocelyn: It's certainly a major achievement, demonstrating robust detection over an unprecedentedly large sky footprint for this population.

Vera: We should all take comfort that even with the limitations of the current data, we have established a clear pathway for future detailed cosmological analyses using these magnificent LBGs.

Subrahmanyan: This sets us up well for next-generation surveys like LSST and DESI-II to build upon this foundation.

Jocelyn: I'm really looking forward to seeing how these findings are utilized in the next great survey of the universe!

Vera: Let's thank all the authors of "Assessing the large-Scale Angular Clustering of UNIONS Lyman Break Galaxies via Cross-Correlations" for their meticulous work.

astro-ph.CO

Submitted: 2026-07-10

Updated: 2026-08-25

Comments: 20 pages, 19 figures, to be submitted to The Open Journal of Astrophysics

Code: https://github.com/LSSTDESC/NaMaster

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 80/100

The gist: The study investigates the feasibility of using LBGs selected from the Ultraviolet Near Infrared Optical Northern Survey (UNIONS) multi-band photometric catalog to probe large-scale structure using

Key concepts

Lyman Break Galaxies (LBGs)
These are high-redshift galaxies used in the survey. They are being studied as a reliable cosmological probe at redshift z~2.5, helping scientists understand structure growth and primordial non-Gaussianity.
Auto-clustering
This is the process of measuring how these specific galaxies cluster with each other (their own clustering). The study found this method unreliable due to pervasive image artifacts across the UNIONS footprint.
Cross-Correlation
A method used to bypass localized imaging problems. It involves measuring the correlation between the LBG distribution and external datasets, such as Planck CMB lensing or DESI/Quaia quasars.
'Namaster' Code
A specific technique mentioned in the methodology designed to remove systematic power on large scales, serving as a sophisticated way to correct instrumental biases within the data.

Terminology

Summary

The following is a detailed summary of the scientific paper:

Lyman-break galaxies (LBGs), which are "powerful tracers of large-scale structure at redshifts z > 2," are valuable cosmological probes used to map the high-redshift, matter-dominated Universe. The study investigates the feasibility of using LBGs selected from the Ultraviolet Near Infrared Optical Northern Survey (UNIONS) multi-band photometric catalog to probe large-scale structure using two-point statistics.

The primary objective was to evaluate the feasibility of measuring large-scale clustering of UNIONS-selected LBGs and assess their potential for cosmological analysis. The authors found that spatially varying imaging systematics, driven by variations in PSF depth, seeing across the UNIONS footprint, limit robust measurements of the LBG auto-angular power spectrum on large scales. This limitation is a common challenge in photometric studies of faint galaxy samples close to survey depth.

Due to these limitations with the auto-spectrum, the study shifts focus to cross-correlation measurements with external tracers, specifically utilizing the Planck CMB lensing convergence map and quasars from DESI DR1 and Quaia, which are described as being less sensitive to the angular imaging systematics.

The methodology involved several stages of analysis:

  1. Characterization: The the authors characterized the LBG density maps, noting that large-scale fluctuations imprint the LBG distribution, revealing the presence of non-astrophysical imaging systematics.

  2. Mitigation: To address these systematics, techniques such as linear regression and de-projection were employed. The namaster code was used to model the observed overdensity (obs) as a linear combination of the true underlying density field (delta LBG) and a set of systematic templates (t i). The corrected overdensity is expressed as:

delta LBG(theta) = delta LBG(theta) - t(theta) F ij d theta' t j(theta')

  1. Validation: The performance of the contamination-decontamination procedure was validated using simulated datasets. The authors found that cross-spectra are robust against variations in contamination amplitude, whereas auto-spectra are more sensitive.

The results demonstrated that:

  • The LBG auto-spectrum (both before and after correction) shows a high clustering amplitude and is severely affected by the systematic variations, making it unsuitable for reliable cosmological interpretation.

  • In contrast, the cross-correlation signal between UNIONS LBGs and external datasets (such as the Planck CMB lensing map) can be measured more robustly than the auto-spectrum.

  • The LBG-CMB lensing cross-power spectrum was found to have an amplitude consistent with theoretical predictions.

In conclusion, the study establishes that UNIONS-selected LBGs are reliable tracers for cross-correlation cosmology at z about 2.5, while simultaneously highlighting that cross-correlation techniques [are] a powerful and robust avenue for extracting cosmological information from photometric high-redshift galaxy samples in the presence of complex imaging systematics. However, the authors also noted that residual angular systematics continue to play a non-negligible role in shaping the statistical uncertainties of the cross-correlation measurements, which directly degrades the constraining power on local PNG.

Improvements for AI systems

Improvements to AI Data Analysis Systems (Focus on Cosmological Inference)

1. Dynamic Statistical Method Selection:

The improved AI system will implement a real-time decision tree that evaluates the characteristics of any given photometric survey's angular distribution before selecting a statistical methodology.

  • What the improved AI can do: Automatically flag when an auto-angular power spectrum (C XX) is compromised by spatially varying imaging systematics (e.g., variations in PSF depth, seeing, or stellar density), and then automatically pivot to prioritize cross-correlation methods (C XY) with external tracers (CMB lensing or QSO datasets). This prevents the use of unreliable auto-spectra for cosmological inference.

2. Automated Non-Linear Systematics Mapping:

The AI will integrate advanced machine learning techniques, specifically Random Forest (RF) regression, to model the relationship between observed LBG overdensity and systematic environmental factors.

  • What the improved AI can do: Instead of relying on linear deprojection methods (namaster), the system can accurately extract a complex, non-linear systematic weight map (w sys) from raw data (mimicking regressis). This allows it to correct for subtle, complex dependencies—such as those caused by local background noise or specific survey tiling patterns—that would be missed by standard linear models.

3. Robust Uncertainty Quantification via Empirical Covariance:

The AI will move beyond the simplistic assumption of Gaussian-only covariance for estimating statistical errors in angular power spectra.

  • What the improved AI can do: The system will automatically calculate and incorporate a Jackknife covariance matrix (Cov(C)) into its Bayesian inference routines. This allows it to accurately quantify the excess variance introduced by residual, uncorrected angular systematics at large scales, providing a far more conservative and accurate constraint on parameters like local primordial non-Gaussianity (f NL) than Gaussian-only estimates.

4. Cross-Correlation Signal Validation and Verification:

The AI will incorporate a simulation framework to validate the robustness of cross-correlation signals against known systematic contamination.

  • What the improved AI can do: It can generate contaminated mock datasets (using techniques like Awan et al. 2025) and test whether the de-contamination pipeline (namaster or regressis) successfully recovers the original input power spectrum (C gg). This allows it to provide a confidence score for the retrieved cosmological signal, guaranteeing that any observed clustering amplitude is consistent with theoretical expectations and not an artifact of uncorrected survey inhomogeneity.

5. High-Redshift Tracing and Bias Modeling:

The AI will manage complex bias models simultaneously within its analysis framework to account for the different characteristics of various tracers.

  • What the improved AI can do: It can simultaneously fit parameters for both the LBG sample (using photometric redshift distributions) and external QSO/CMB lensing fields, while applying appropriate bias prescriptions (e.g., Wilson & White 2019). This allows it to rigorously test whether the cross-correlation signal is consistent with a unified physical model of structure formation, even across different tracer populations.

Abstract

Lyman-break galaxies (LBGs), selected via the strong spectral break blueward of the Lyman limit, are powerful tracers of large-scale structure at redshifts z>2. In this work, we assess the feasibility of using LBGs selected from the Ultraviolet Near Infrared Optical Northern Survey (UNIONS) multi-band photometric catalog as cosmological probes of the high-redshift Universe using two-point statistics. We demonstrate that spatially varying imaging systematics, driven by variations in PSF depth, seeing across the UNIONS footprint, limit robust measurements of the LBG auto-angular power spectrum on large scales, even after correcting the LBG field with linear or non-linear mitigation techniques. This study shows that clustering analyses of faint galaxy samples close to survey depth are challenging. We therefore turn to cross-correlation measurements with external tracers, in particular the Planck CMB lensing convergence and quasars from DESI DR1 and Quaia, which are less sensitive to the angular imaging systematics. Using both data and mock catalogues, we demonstrate that the LBG--CMB lensing cross-power spectrum can be measured more robustly than the auto-spectrum, with an amplitude consistent with theoretical predictions. Residual systematics primarily manifest as excess variance at large angular scales, without introducing a significant bias in the recovered signal. Taken together, these results establish UNIONS-selected LBGs as reliable tracers for cross-correlation cosmology at z about 2.5, and highlight cross-correlation techniques as a powerful and robust avenue for extracting cosmological information from photometric high-redshift galaxy samples in the presence of complex imaging systematics.

Sources

Related papers