Skip to main content
eScholarship
Open Access Publications from the University of California

UC Riverside

UC Riverside Electronic Theses and Dissertations bannerUC Riverside

A Statistical Modeling Approach to Selection and Study of Galaxies at Different Phases of their Star-Formation Activity at High Redshift

Creative Commons 'BY' version 4.0 license
Abstract

This thesis focuses on the selection and study of galaxies based on their star formation activity at high redshift. I use multiple selection techniques from the traditional color selection to a Bayesian model averaging approach with Bayesian SED fitting to select the massive quiescent galaxies at $z \geq 3$ in the Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS). I compare predictions from empirical to the latest cosmological hydrodynamical simulations and find that number and stellar mass density are higher than predictions. I estimate their halo mass using abundance matching, which results in massive enough halos that shock heating for some fraction of the gas can explain part of their quenching process. But as cold streams are expected to be significant even for massive halos at these redshifts, other quenching mechanisms such as AGN feedback must have the dominant role. I then develop models for selecting these objects using a statistical learning approach to allow a robust and computationally efficient selection in the upcoming extensive surveys. I train and validate different methods using the mock catalog from the semi-analytic models for the CANDELS. Many of these techniques outperform the generic SED-fitting approach applied on the large catalogs and make more robust samples in terms of completeness and purity. Finally, I build a probabilistic model for jointly describing the galaxies' stellar mass, star-formation rate, and local density contrast, using a mixture model while accounting for different sources of uncertainties. I find that the effect of the environment on the prediction of a galaxy's star-formation activity is different when moving towards higher redshifts. The impact of the environment on the odds of being quiescent versus star-forming has small interaction with stellar mass at low redshift but shows strong interaction at high redshift ($> 1$) such that the effect of the environment is larger for more massive galaxies. The effect of the mass also depends on the environment and becomes larger in a denser environment. This is partly consistent with the picture that galaxies with halted cosmological gas accretion can become quiescent by depleting their gas reservoir through star-formation and outflows.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View