# High-redshift LBG selection from broadband and wide photometric surveys using a Random Forest algorithm

**Authors:** C. Payerne, W. d'Assignies Doumerg, C. Yèche, V. Ruhlmann-Kleider, A. Raichoor, D. Lang, J. N. Aguilar, S. Ahlen, D. Bianchi, D. Brooks, T. Claybaugh, S. Cole, A. de la Macorra, B. Dey, P. Doel, A. Font-Ribera, J. E. Forero-Romero, S. Gontcho A Gontcho, G. Gutierrez, K. Honscheid  
**arXiv:** [2410.08062](https://arxiv.org/abs/2410.08062)  
**Date:** 2024-10-10  
**Summary page:** https://www.summarizepaper.com/en/paper/2410.08062v1/

## Key points

- Authors explore selection of high-redshift Lyman-Break Galaxies (LBG) using current and upcoming broadband wide photometric surveys
- Random Forest algorithm used to identify LBGs in surveys like UNIONS or Vera C. Rubin LSST
- Research conducted within framework of future large-scale structure spectroscopic surveys such as DESI-II
- Deep imaging data from HSC and CLAUDS on COSMOS and XMM-LSS fields leveraged for analysis
- Methodology yields LBG density of 873 deg$^{-2}$ for targets with redshifts greater than 2 post-spectroscopic validation with DESI
- Secure spectroscopic sample generated with average redshift of 3 for DESI-II forecasts
- Precise measurements of Alcock-Paczynski parameters $\alpha_\perp$ and $\alpha_{\parallel}$ at levels of 0.7% and 1%, respectively, within redshift range of $2.6 < z < 3.2$
- Prediction of dark energy fraction measurement at 2% precision
- Estimate uncertainty in local non-Gaussianity with predicted value close to $\sigma_{f_{\rm NL}} \approx 7$, comparable to best precision achieved by Planck

## Summary

, , , , 

In this paper, the authors explore the selection of high-redshift Lyman-Break Galaxies (LBG) using current and upcoming broadband wide photometric surveys. They employ a Random Forest algorithm to identify these galaxies in surveys like UNIONS or the Vera C. Rubin LSST. This research is conducted within the framework of future large-scale structure spectroscopic surveys such as DESI-II, slated to commence around 2029. The researchers leverage deep imaging data from HSC and CLAUDS on the COSMOS and XMM-LSS fields to carry out their analysis. By degrading the $u, g, r, i$, and $z$ bands to UNIONS depth, they predict how well LBGs with similar image quality can be selected using this approach. The Random Forest algorithm is trained using these bands to classify LBGs falling within a redshift range of $2.5 < z < 3.5$. With a target density budget of $1,100$ deg$^{-2}$ fixed, the methodology yields a density of $873$ deg$^{-2}$ for targets with redshifts greater than 2 and confirms a density of $493$ deg$^{-2}$ for LBGs post-spectroscopic validation with DESI. A dedicated spectroscopic observation campaign involving 1,000 targets with DESI on the COSMOS field was carried out to test this UNIONS-like selection process. This resulted in a secure spectroscopic sample showcasing an average redshift of 3 which was then used to generate forecasts for DESI-II under an assumption of sky coverage spanning 5,000 deg$^2$. The authors predict uncertainties in Alcock-Paczynski parameters $\alpha_\perp$ and $\alpha_{\parallel}$ at levels of 0.7$\%$ and 1$\%$, respectively, within the redshift range of $2.6 < z < 3.2$. This leads to a precise measurement of dark energy fraction amounting to 2%. Additionally, estimates are made regarding uncertainty in local non-Gaussianity with a predicted value close to $\sigma_{f_{\rm NL}} \approx 7$, comparable to current best precision achieved by Planck. Overall, this study showcases promising results in utilizing advanced algorithms and cutting-edge survey data for efficient selection and characterization of high-redshift Lyman-Break Galaxies in preparation for future large-scale structure spectroscopic surveys like DESI-II.