Galaxy evolution: A new version of the Besançon Galaxy Model constrained with Tycho data

Abstract

The understanding of the origin and evolution of the Milky Way is one of the primary goals of the Gaia mission (ESA, launch autumn 2013). In order to study and analyse fully the Gaia data it will be useful to have a Galaxy model able to test various hypothesis and scenarios of galaxy formation and evolution. Kinematic and star count data, together with the physical parameters of the stars - ages and metallicities-, will allow to characterize our galaxy's populations and, from that, the overall Galactic gravitational potential. One of the promising procedures to reach such goal is to optimize the present Population Synthesis models (Robin et al. (2003)) by fitting, through robust statistical techniques, the large and small scale structure and kinematics parameters that best will reproduce Gaia data. This PhD thesis was focused on the optimization of the structure parameters of the Milky Way Galactic disc. We improved the Besançon Galaxy Model and then by comparing the simulations to real data studied the process of Galaxy evolution.

The Besançon Galaxy Model is a stellar population synthesis model, built over the last two decades in Besançon (Robin and Crézé(1986); Robin et al. (2003)). Until now the star production process in that model was based on the drawing from the so called Hess diagrams. Each Galaxy population had one such a diagram, which was calculated once given a particular Initial Mass Function (IMF), Star Formation Rate (SFR), evolutionary tracks and age-metallicity relation and since then remained fixed in the model. As that feature was not enabling to test any other scenario of Galaxy evolution, because none of the evolutionary parameters could be modified, it was one of the biggest weaknesses of the model. It has served us as a motivation to dedicate this PhD project to the construction of a new version of the model, which would be able to handle variations of the SFR, IMF, evolutionary tracks, atmosphere models among others. When the evolutionary parameters are changed one must repeat the process of accomplishing the dynamical self-consistency of the model as described in Bienayme et al. (1987). For that we have recalculated the Galactic gravitational potential for all new evolutionary scenarios, which have been tested. The second very important improvement of the model, which is delivered in this thesis, is the implementation of the stellar binarity. That is, the new version of Besançon Galaxy Model presented here is not any more a single star generator, but it considers binary systems maintaining constraints on the local mass density. This is an important change since binaries can account for about 50 % of the total stellar content of the Milky Way.

Once the tool was developed we tested several possible combinations of IMF and SFR in the Solar Neighborhood and identified those which best reproduce the Local Luminosity Function and Tycho-2 data. We have accomplished an unprecedented task using the new version of the model, namely we have performed the whole sky comparisons for a magnitude limited sample in order to study the bright stars. The Tycho-2 catalogue turned out to be an ideal sample for that task due to its two important advantages, the homogeneity and completeness until VT ~ 11 mag.

Different techniques and strategies were designed and applied when comparing the simulated and the real data. We have looked at small and specific Galactic directions and also performed general comparisons with a global sky coverage. In order to increase the efficiency of numerous simulations and comparisons, a processing pipeline based on C, Java and scripting programming languages has been developed and applied. It is a fully automated, portable and robust tool, allowing to split the work across several computational units.

La misión Gaia (ESA, 2013) revolucionará el conocimiento sobre el origen y la evolución de nuestra Galaxia. Una óptima explotación científica de sus datos requiere disponer de modelos que permitan contrastar hipótesis y escenarios sobre estos procesos de formación. En esta tesis hemos optimizado el modelo de síntesis de poblaciones estelares de Besançon, ampliamente utilizado por la comunidad internacional, centrándonos en la componente del disco delgado. Hemos diseñado, desarrollado, implementado y testeado una nueva estructura de generación de las estrellas que permite encontrar la mejor combinación de función inicial de masa (IMF) y ritmo de formación estelar (SFR) que ajusta a las observaciones. El código permite imponer la autoconsistencia dinámica, recalculando el potencial galáctico para cada nuevo escenario de evolución. También, por primera vez, se generan sistemas binarios bajo esta consistencia dinámica, marcada por la función de luminosidad observada en el entorno solar. Esta, junto con el catálogo Tycho, han sido los dos ingredientes observacionales clave para el ajuste entre modelo y observación. También, por primera vez, hemos conseguido un ajuste aceptable a los recuentos estelares de todo el cielo hasta V=11. Se han evaluado con rigor los efectos en los recuentos estelares derivados del uso de los modelos de atmosfera, de evolución estelar y de extinción interestelar así como de parámetros tan críticos como la masa dinámica del sistema galáctico. El ajuste de estos ingredientes usando el catálogo Tycho nos ha permitido confirmar, de una vez por todas, que la SFR en el disco galáctico no ha sido constante sino decreciente desde los inicios de la formación de esta estructura. En conclusión, esta tesis proporciona un nuevo código, optimizado y flexible en el uso de los ingredientes básicos, en el que se ha realizado una rigurosa evaluación y actualización de los ingredientes que lo componen.