Resumen de Astrostatistics for luminosity calibration in the gaia era
Max J. Palmer
The European Space Agency astrometric satellite \emph{Gaia} was launched in December 2013. Having completed its commissioning phase, Gaia is transmitting data to Earth that will lead to the creation of an astrometric catalogue of more than 10^9 celestial objects. This catalogue will contain astrometric measurements, including parallaxes, with an expected precision in the 10-300 muas range. Due to the very large quantity of data which will be produced, it is essential to begin preparing for the eventual use of the Gaia catalogue. In this thesis we summarise the expected contents of the Gaia end-of-mission catalogue using state of the art simulations from the Gaia Data Processing and Analysis Consortium.
To prepare for the use of the Gaia catalogue, we define a set of statistical methods for the correct use of trigonometric parallax data. Direct use of the distance obtained from the inverse of the parallax is problematic, due to the possible effect of numerous known statistical biases and selection effects. We provide a set of methods which utilise parallax information correctly, in order to calibrate the absolute luminosity of several objects of interest: normal stars, variable stars, open clusters, and the Magellanic Clouds. The methods make extensive use of statistical modelling and Bayesian statistics in order to treat the data correctly and without bias. The methods have been tested using simulated Gaia data, and are ready to be applied to the real Gaia data after release.
In the case of Cepheid and RR-Lyrae variable stars, we present a method for fitting the period-luminosity relation either with parallaxes, or without. For the parallax free case, the method has been applied to real data in order to calibrate the period-luminosity-metallicity relation for RR-Lyraes in the LMC. In the case of open clusters, the method has been applied to Hipparcos data in order to estimate the distance to the Pleiades and Hyades. We contribute to the long-standing Pleiades distance debate through an analysis of the effect of correlated errors in the original (1997) and new (2007) hipparcos reductions. With the possibility of the correlated errors affecting the maximum precision obtainable with Gaia data, we study the expected effect of error correlations in Gaia.
