Resumo:
We present in this work preliminar results on galaxy evolution studies. These studies aim to examine issues such as the evolution in density and luminosity as a function of the age of the Universe in redshifts up to 1 approximately. The main goal of this work was the implementation of a full methodology, since the treatment of the photometric data, leading to the estimation of the Luminosity Function of Galaxies (LF). In the near future, this methodology will be applied on data from Dark Energy Survey (DES) annual releases. Initially, a pipeline to estimate the galaxy Luminosity Function under the DES Scientific Portal scope was implemented. Given the large volume of data, tools to keep the source of the data, distribute catalogs and share results with other members of the collaboration becomes an expedient, more than helpful - necessary. Since DES essentially produces photometric data, we need to estimate distances to galaxies using photometric redshifts. A great effort was made for the creation of scientific workflows that make the training and validation samples for several of these algorithms. Furthermore it is necessary to apply the algorithms to the samples for which one desires to determine the photo-z's -- called photometric sample. We present a performance analysis of these codes for simulated and Science Verification phase of DES by checking the impact of these results in the determination of Luminosity Function. We examined the impact of different criteria for sample selection, such as limits on the signal-to-noise ratio and the magnitude limit of the sample in the range of redshifts used, for the calculation of photo-z's. We also tested the performance of the algorithms without u-band, which is not present in DES. All these tests are used to assess the impact in determining the Luminosity Function for different redshift bins, so one can examine the behavior of the density and magnitude characteristics of the Universe as a function of age, one of the goals of our study on the evolution of galaxies. We learned from this work that the quality of the data is as important as the methodology adopted, and the cleaning of the sample, based on signalto noise ratio has been an important procedure, having its impact characterized by improvement on photo-z metrics. The combination of the best accuracy obtained with empirical codes and the possibility of obtaining other results, in addition to photo-zs, provided by template-fitting methods seems to be a valuable approach to the study of evolution of galaxies through the Luminosity FunctioPhotometric Redshifts; Luminosity Function of Galaxies; Galaxy Evolution.