Mapping Galactic Bubbles, Shells, and Clouds with Persistent Homology: Topology-informed Structure Identification in 2D and 3D Data

Feedback-driven bubbles are increasingly recognized as playing a critical role in regulating star formation and shaping the structure of galaxies. Studies of bubbles in the Solar Neighborhood and in external galaxies have suggested that star and cloud formation is frequently situated at the edges of expanding cavities, and that the ISM may be riddled with colliding and merging voids. As the amount of data we possess about the structure of the Milky Way exceeds quantities that can be parsed efficiently by eye, robust computational methods are needed for the identification and hierarchical mapping of bubbles, shells, and clouds. We present a new method for the topological identification and geometric mapping of diffuse structures in 2D images and 3D volumes, informed by the technique of persistent homology from the field of computational algebraic topology. We apply this method to a new pc-resolution 3D dust map of the Solar Neighborhood (Edenhofer et al. 2024) and map the geometries of both candidate feedback-driven bubbles and diffuse atomic and molecular clouds within 1.2 kpc from the Sun. We discuss the role that bubbles may have played in shaping the structure of the Milky Way and promoting star formation in the Solar Neighborhood. Our topological structure identification method will be publicly released as a python package “perch” and can in the future be used to analyze the structure and morphology of 2D and 3D surveys of the Milky Way and external galaxies.