Multi-method approach to membership selection in ultra-faint dwarf galaxies

Ultra-faint dwarf galaxies (UFDs) are systems that are primarily characterised by old and metal-poor stellar populations, with ages and metallicities comparable to those of the oldest and most metal-poor stars in the Galactic halo. This scenario raises the possibility of a evolutionary connection between these systems and the formation process of the Milky Way (MW) halo. In this context, RR Lyrae stars (RRLs) are among the most powerful tracers of old and metal-poor stellar populations such as those observed in UFDs. Using data from the third data release (DR3) of the European Space Agency mission Gaia , we investigated the outskirts of six UFDs surrounding the MW to the purpose of consolidating their RRLs content and possibly identifying new RRL members of these systems. Two complementary methodologies were applied to the Gaia DR3 data. The first is a classical approach based on a proper motion (PM) analysis, period-Wesenheit-metallicity relation for RRLs, and colour-magnitude diagram. The second is a statistical approach using machine learning techniques to associate RRLs across the sky with their host galaxy based on astrometric and photometric properties. The analysis allowed us to consolidate the census of member variables within each UFD: 25 RRLs in Boötes I, 45 in Boötes III, 3 in Carina II, 12 in Coma Berenices, 1 in Sagittarius II, and 8 in Ursa Major I. Compared with the results in the literature, we managed to increase by one unit the RRL population of Boötes I, although we did not confirm three RRL members from literature. We confirmed 20 more for Boötes III, with a loss of 7 RRLs; 9 units for Coma Berenices; and 4 candidates were not confirmed for Sagittarius II; finally, we recovered the same variables for Carina II and Ursa Major I. We also computed the average PM of the galaxies by using the member stars confirmed via spectroscopic analysis.