Home > Rationale

Rationale

‘To contribute usefully to the scientific progress, one must sometimes not disdain from undertaking simple verifications’ (Foucault, 1847).

Despite the precautions taken when building the Gaia data processing, completely avoiding errors in a one billion+ source catalogue, with many intricate data for each object, is an impossible task. Before the publication of the Gaia archive, an in-depth validation of its contents has thus to be undertaken by the Gaia DPAC. It relies on methods and tools developed in the CU9 Work Package 940 in charge of the scientific validation of the catalogue.

Ensuring there are no systematic errors will lead to check for suspect spatial (or epoch, colour, magnitude, ...) variations, which could be due to observational or processing artefacts. The validation process will thus address the statistical distribution of the astrometric parameters, typically the unbiasedness of e.g. the parallaxes and proper motions, which are the main outcomes of the mission, and which may be subject to many perturbations. It will also scrutinize the photometric and spectroscopic data, assisted by the important classification task performed in Gaia, and, when available, external data.

Most of the validation work is being done with 8 work packages:

  • WP942: Internal consistency and more complex scenarios
  • WP943: Comparing models with data.
  • WP944: Confrontation with external archives.
  • WP945: Statistical tools
  • WP946: Time series and variability
  • WP947: Clusters as validation tools
  • WP948: Solar System Objects

Besides the monthly telecons with the work package managers, it has been decided to have two plenary validation meetings per year.

  • The first meeting in Vienna in July 2014 described the work done in the various Work Packages.
  • The CU9(940) and GENIUS (T5) workshop in Meudon on March 2-4, 2015 rather intended to enter more into the details of all the validation procedures, by familiarising with the validation development and runtime environments, studying and improving current methods, running tests on existing or simulated catalogues.
  • The Barcelona meeting on Sept. 9, 2015 allowed to analyse the first validation results, either done in CU3 or in the validation group.

This Paris meeting, which is the fourth plenary, will:

  • present the validation work done within the CUs and CU9 Work Packages
  • discuss about the problems found, what should be: accepted (i.e. documented) or corrected (objects deleted, e.g. artefacts)
  • envision possible further tests
  • prepare for a possible re-run (i.e. re-doing validation after corrections)