Accurate photometry and photometric redshifts in cosmological surveys

Laburpena

Due to the finite speed of light, which imposes any physical event to not instantaneously propagate though space-time, the history of the universe can be traced back. This very fact makes possible to observe the universe as it was in other epochs and so understand the physical processes that, millions of years ago, dictated the distribution, formation and evolution of the galaxies we observe today. The expansion of the universe stretches the electromagnetic radiation emitted by extragalactic sources while traveling through the space-time. This quantifiable effect, commonly known as cosmological redshift, has become a widely-accepted indicator to measure cosmological distances, and so understand when (in terms of the cosmic time) the light that we register today in our observatories departed originally from the galaxies. Interpreting this spectral shifting as a physical distance certainly requires the support of a theoretical framework. The theory of General Relativity (or theory of gravity) has played a fundamental role in describing the present-day universe, successfully explaining the expansion history of the universe in terms of the energy-mass content and the geometry of the space-time, through the evolution of the scale factor. During the last decades, cosmology, the science of the universe as a whole, has experienced a tremendous progress. Today, the Lambda-Cold-Dark-Matter (¿CDM) scenario is widely accepted as the standard model of cosmology describing not only the evolution of the universe but also the different constituents that may have populated it over most its cosmic time. The different ingredients and their proportions are expressed through the cosmological parameters, which represents the real physical quantities that can be retrieved by observations. One of the reasons why modern cosmology has achieved a solid world-model is due to the carefully designed observational programs that have systematically surveyed the Universe. This way, the systematic combination of multiple studies (such as the Cosmic Microwave Background Radiation (CMB), the Dark Matter content and distribution on Massive Galaxy Clusters, the Dark Energy equation of state via distant Supernovae or measuring the Baryonic Acoustic Oscillations (BAO), the re-ionization epoch examining high-z QSO spectra, the Big Bang nucleosynthesis of the primordial abundances of light elements, the galaxy formation and evolution or the large scale structure (LSST) of galaxies, among others) has made possible that most cosmological parameters are known to a few percentage accuracy now. The main research activities of this PhD thesis have been focused on the acquisition of accurate photometric redshift catalogues for two cosmological surveys: the ALHAMBRA-survey (Moles07) and the CLASH-survey (Postman12). Facing all the complexity behind the pipelines. To pass from the astronomical images to the scientific data which will eventually be used to derive any cosmological analysis. As several times discussed during the manuscript, performing complete, unbiased, homogeneous and reliable measurements for the observed galaxies in the sky is a delicate task. Although the two projects involved in this work shared exactly the same goal and methodology (estimate galaxy redshifts via multiband photometry), it was (at certain points) impractical to freely apply the same analytical tools from one dataset to another. The kind of problems associated to each survey made the design of the pipelines quite complex, rather than a repetitive task, emphasizing the role of observational astronomy in the neat progress of modern cosmology. After a brief introduction to the Cosmological Model (Chapter1) and a quick discussion of the role of redshift surveys and the advantages and drawbacks of the different methodologies used to infer the galaxy redshifts (Chapter2), this manuscript carefully describes (during Chapter 3 & 4) the methodology used to successfully retrieve accurate photometric redshift estimations for both surveys. In particular, this thesis provides a guide on the following analytical processes: * How to compute a reliable PSF-corrected multi-band photometry using the ColorPro software (Coe et al. 06, Molino et al. 2014) to meet the specification of the survey. * How to generate PSF-models (as required by ColorPro) per each individual image, by carefully selecting hundredths of well-isolated and good photometric stars to assure the fidelity of the final models. * How a new approach to generate broadband images, as a combination of individual bands, was developed and implemented. Serving for ALHAMBRA to create synthetic HST/ACS F814W detection images, defining a constant, homogeneous and comparable window for all the ALHAMBRA fields with other projects like the COSMOS-survey. * How to design set of simulations to assure the goodness of ColorPro retrieving precise photometry across images with varied PSF. * How to mask out saturated stars, stellar spikes, ghosts or simply damaged areas through the images improving as much the source detection efficiency as the background subtraction process. * How to decontaminate extragalactic sources from field stars, using a statistical classification method * How to recalibrate the photometric zero-point (PZP) calibrations using the colors of the emission line galaxies predicted by photometric redshifts. * How to use and configure the Bayesian Photometric Redshift (BPZ) software, exploit the whole information contained in the complete redshifts probability distribution function (P(z)). * How to quantify the photometric redshifts accuracy in terms of the magnitude, the redshift range or the meaningful ODDs parameter. * How to identify potential AGN and QSO candidates using the BPZ software. * How to perform accurate photometry on massive galaxy clusters, where the contamination of the Intra Cluster Light (ICL) disrupts the colors of the galaxy making impracticable to retrieve accurate photometry. * How to run a set of simulations to estimate the most convenient background-subtraction to generate Background-free images. * How generate a synthetic sample of spectroscopic redshifts galaxies to quantify the photometric bias induced by the ICL. * How to empirically derive reliable photometric uncertainties for the images, considering as much the correlation among pixels as the photometric bias made by SExtractor when estimation ISOphotal apertures. * How to verify that Background-subtraction treatments do not alter the photometric colors derived on Background-free images. * How to retrieve total magnitudes using Background-free detection images, using optimal photometric apertures. * How to calculate and implement realistic upper limits for photometric redshifts, avoiding problems of misclassification of galaxies at low redshifts to high redshifts. * How to verify that a PSF-corrected photometry is mandatory when combining images from different detectors or different qualities. * How to verify the improvement of photometric redshift estimations after applying the whole procedure in comparison to standard treatments. Finally, during Chapter 5&6, the participation (and the most relevant publications) in the "CLASH: High-Redshift Universe" group and the "CLASH: Supernovae" group are briefly explained, where the different analysis and requirement for these two additional projects are fully discussed.