DESI DR2 Reference Mocks: Clustering results from UCHUU ELGs and QSOs

R. Vaisakh, J. Lasker, R. Kehoe, A. Amalbert, N. Khan, E. Fernandez-Garcia, F. Prada, M. S. Wang, J. DeRose, S. Bailey, A. J. Ross, J. Aguilar, S. Ahlen, D. Bianchi, D. Brooks, F. J. Castander, T. Claybaugh, K. S. Dawson, A. de la Macorra, S. Ferraro, J. E. Forero-Romero, E. Gaztanaga, Satya Gontcho A Gontcho, G. Gutierrez, C. Hahn, M. Ishak, R. Joyce, S. Juneau, T. Kisner, A. Kremin, C. Lamman, M. Landriau, M. E. Levi, M. Manera, R. Miquel, A. D. Myers, S. Nadathur, W. J. Percival, I. Perez-Rafols, G. Rossi, E. Sanchez, D. Schlegel, H. Seo, G. Tarle, B. A. Weaver, R. Zhou, H. Zou

First listed 2026-06-30 | Last updated 2026-06-26

Abstract

High-redshift galaxy clustering provides a powerful probe of the growth of structure, testing models of dark matter, dark energy, and galaxy formation during the epoch when the Universe was rapidly evolving. Emission line galaxies (ELGs) and quasars (QSOs) are used as tracers of dark matter by the Dark Energy Spectroscopic Instrument (DESI) to probe this redshift regime. We present results from ELG and QSO mock catalogs created from the Uchuu N-body simulation and tuned to DESI Data Release 2 (DR2) clustering. Employing a modified subhalo abundance matching (SHAM) technique, we populate Uchuu halos and subhalos with QSOs between 0.8 < z < 2.1. For ELGs, we modify this method to select satellite galaxies with low velocities relative to their associated central halos, and populate a separate set of Uchuu halos and subhalos with ELGs between 0.8 < z < 1.6. In this paper, we reproduce the redshift evolution of number density and clustering statistics across the fitted range of scales. We also measure the large-scale clustering bias of both the data and mock samples. These results improve simulated lightcone construction from cosmological models and enhance our understanding of the galaxy-halo connection.

Short digest

This paper builds DESI DR2 reference mocks for emission-line galaxies and quasars by populating the Uchuu N-body simulation with a modified SHAM model, fitting ELGs over 0.8 < z < 1.6 and QSOs over 0.8 < z < 2.1. The key empirical tweak is tracer-specific: QSO clustering is reproduced with a redshift-dependent satellite fraction and a fixed characteristic velocity scale, while ELGs require selecting satellites with low velocities relative to their centrals and an increasing V_mean with redshift. With those choices, the mocks recover the observed redshift evolution of number density, the monopole across the fitted 3 to 70 h^-1 Mpc range, and also give reasonable quadrupole, projected-correlation, and power-spectrum behavior without being directly fit to all of them. The payoff is a high-fidelity lightcone set for DESI clustering analyses that also yields large-scale bias trends and halo-occupation constraints, sharpening how ELGs and QSOs map onto dark matter halos.

Key figures to inspect

• Figure 1. Use this figure to show that the mocks are not just tuned to clustering amplitude but also track the observed comoving number-density evolution of both ELGs and QSOs across 0.8 < z < 2.1. That redshift-dependent abundance match is foundational for any later claim that the lightcones are realistic survey references rather than single-epoch fits.
• Figure 3. This is the clearest summary of the paper’s modified SHAM ingredients. It isolates the main physical and phenomenological result of the modeling: QSOs prefer an increasing satellite fraction with redshift at fixed V_mean, while ELGs are better described by a fixed satellite fraction and a redshift-rising V_mean, including the low-relative-velocity satellite selection that distinguishes the ELG treatment.
• Figure 4. Choose this as the main evidence figure because the abstract’s central claim is that the mocks reproduce DESI DR2 clustering over the fitted range. The monopole comparison directly demonstrates agreement between data and the best-fit modified SHAM models for both tracers, making it the most important validation panel in the paper.
• Figure 14. This figure connects the successful clustering fit back to the galaxy-halo connection, which the abstract explicitly frames as a key outcome. The separate central and satellite occupation curves for ELGs and QSOs make clear how the two tracers inhabit halos differently and why the modified SHAM prescriptions are physically informative rather than purely empirical.
• Figure 15. Use this as the synthesis figure for the paper’s large-scale conclusions. It shows the redshift evolution of bias measured in both DESI data and Uchuu mocks, confirming that the calibrated lightcones capture the monotonic rise in tracer bias and therefore provide a useful basis for future DESI cosmological analyses.