A hierarchical clustering method for quantifying satellite abundance

TL;DR

This paper introduces a statistical hierarchical clustering method to estimate satellite galaxy abundance around field galaxies using incomplete redshift data, validated with COSMOS survey data.

Contribution

The paper presents a novel probabilistic clustering method that estimates satellite abundance without identifying individual satellites or centrals, suitable for incomplete redshift samples.

Findings

Consistent satellite abundance results with previous studies

Extended satellite luminosity function over a broader mass range

Systematic uncertainties smaller than random errors

Abstract

We present a new method for quantifying the abundance of satellites around field galaxies and in groups. The method is designed to work with samples, such as local photometric redshift catalogues, that do not have full spectroscopic coverage, but for which some redshift or distance information is available. It consists of identifying the galaxies most likely to be centrals, and using the clustering signal around them as a template to iteratively decompose the full population into satellite and central populations. In that sense it is similar to performing crowded-field photometry, after having first used isolated stars to determine the point spread function of the image. The method does not identify individual satellites or centrals conclusively, but assigns a probability to each galaxy of being one or the other. Averaged over a large sample, it provides a statistical estimate of satellite abundance, even in crowded fields with large redshift uncertainties. We test the method using data from the COSMOS field, which includes a large set of local objects with accurate photometric redshifts. We measure satellite abundance as a function of central stellar or halo mass, as well as the satellite luminosity function, and find results consistent with previous studies, but extending over a broader range of central masses. We also consider a number of possible systematic uncertainties in the method, and show that they are generally smaller than our random errors. Having presented the method in this paper, we will use it to study the properties of the satellite populations in a forthcoming one.