Multi-Band Galaxy surveys

Abstract

This doctorate thesis covers the use of recent astronomical surveys and statistical techniques to answer questions about the composition of galaxies, by following two different perspectives: the instruments needed to gather the data and the development of a statistical tool to turn the observed data into physical properties of galaxies. We start by giving some background contextualization about the largest extragalactic surveys to date, and how they influenced the three surveys studied in this thesis: the S-PLUS survey, the ALHAMBRA survey, and the J-PAS survey. The first half of this work is dedicated to the S-PLUS survey, a Local Universe survey that started at the beginning of 2017 and is designed to study nearby galaxies in 12 bands with a 80 cm telescope. Its design, installation, and commissioning occurred in the course of this thesis and both hard and software components of the observatory are described as well how they are projected to map the sky autonomously every night. In the second half of this work, we turn our attention into the data generated by other two multi-band surveys: the ALHAMBRA and JPAS surveys. The ALHAMBRA survey, which completed the map of 3 deg2 in 20 medium band optical filters on 2014, provided the photometric data and redshifts (photo-zs) used in this work. The studies performed here with the ALHAMBRA data will provide guidelines to explore the data of its successor, the J-PAS survey, which starts next year building a map of the whole northern sky in 56 narrow band filters. We show how the problem of the estimation of the photo-z in the Universe is unfolded by the use of Bayesian statistics, where galaxies positions are described not by a single point estimator, but by a probability distribution function (PDF). Within the Bayesian photo-z framework, we develop a method for estimation of galaxy properties using the template-redshift PDFs given by photometric redshift codes as, for example, the BPZ code. We propose a concrete implementation of this method by fitting the photo-z templates with composite stellar populations models, or -models. The output of the proposed implementation is used to, in combination to the ALHAMBRA photo-z PDFs, estimate galaxies stellar properties like the mean ages, extinctions, and stellar masses. The derived stellar mass PDFs were used to estimate the stellar mass function in different redshift bins. The resulting ages, stellar masses, and the stellar mass functions are shown to be systematically inclined to make galaxies bluer and, hence, less massive than the expected. This deficit of massive galaxies in the application of the method is discussed, and the solution proposed relies on changing the photo-z templates to more physical realistic ones while maintaining the so-pursued photo-z precision.