The SDSS-IV eBOSS: emission line galaxy catalogues at z=0.8 and study of systematic errors in the angular clustering

TL;DR

This paper presents two large, systematically corrected catalogs of emission line galaxies at redshift 0.8, designed for cosmological studies, and demonstrates effective removal of observational systematic errors from their angular clustering measurements.

Contribution

The paper introduces the largest ELG catalogs at z=0.8 with systematic error correction methods, enhancing their utility for cosmological analyses.

Findings

Systematic errors can be effectively modeled and removed using multivariate regression.

Fluctuations in imaging zero-points have minor impact on galaxy distribution.

The catalogs exhibit similar bias parameters suitable for large-scale structure studies.

Abstract

We present two wide-field catalogs of photometrically-selected emission line galaxies (ELGs) at z=0.8 covering about 2800 deg^2 over the south galactic cap. The catalogs were obtained using a Fisher discriminant technique described in a companion paper. The two catalogs differ by the imaging used to define the Fisher discriminant: the first catalog includes imaging from the Sloan Digital Sky Survey and the Wide-Field Infrared Survey Explorer, the second also includes information from the South Galactic Cap U-band Sky Survey (SCUSS). Containing respectively 560,045 and 615,601 objects, they represent the largest ELG catalogs available today and were designed for the ELG programme of the extended Baryon Oscillation Spectroscopic Survey (eBOSS). We study potential sources of systematic variation in the angular distribution of the selected ELGs due to fluctuations of the observational parameters. We model the influence of the observational parameters using a multivariate regression and implement a weighting scheme that allows effective removal of all of the systematic errors induced by the observational parameters. We show that fluctuations in the imaging zero-points of the photometric bands have minor impact on the angular distribution of objects in our catalogs. We compute the angular clustering of both catalogs and show that our weighting procedure effectively removes spurious clustering on large scales. We fit a model to the small scale angular clustering, showing that the selections have similar biases of 1.35/D_a(z) and 1.28/D_a(z). Both catalogs are publicly available.