Halo Properties from Observable Measures of Environment: I. Halo and Subhalo Masses

TL;DR

This study uses neural networks to analyze how observable environment measures relate to dark matter halo and subhalo masses, revealing that local galaxy environment and stellar mass are key indicators depending on halo mass.

Contribution

The paper demonstrates that neural networks can effectively predict halo masses from observable environment measures, highlighting the differing importance of local environment based on halo mass scale.

Findings

For high-mass halos, environment within 1 Mpc constrains halo mass.

Below a certain mass, stellar mass alone predicts halo mass effectively.

Neural networks achieved an error of 0.20 dex in halo mass prediction.

Abstract

The stellar mass - halo mass relation provides a strong basis for connecting galaxies to their host dark matter halos in both simulations and observations. Other observable information, such as the density of the local environment, can place further constraints on a given halo's properties. In this paper, we test how the peak masses of dark matter halos and subhalos correlate with observationally-accessible environment measures, using a neural network to extract as much information from the environment as possible. For high mass halos (peak mass $>10^{12.5} M_{\odot}$), the information on halo mass contained in stellar mass - selected galaxy samples is confined to the $\sim$ 1 Mpc region surrounding the halo center. Below this mass threshold, nearly the entirety of the information on halo mass is contained in the galaxy's own stellar mass instead of the neighboring galaxy distribution. The overall root-mean-squared error of the best-performing network was 0.20 dex. When applied to only the central halos within the test data, the same network had an error of 0.17 dex. Our findings suggest that, for the purposes of halo mass inference, both distances to the $k$th nearest neighbor and counts in cells of neighbors in a fixed aperture are similarly effective measurements of the local environment.