Tracing the Morphology-Density Relation with the DESI Bright Galaxy Survey

The morphology-density relation, the observed correlation between a galaxy’s structure and the density of its environment, is a key diagnostic of galaxy evolution. We trace its development with the Dark Energy Spectroscopic Instrument (DESI) Bright Galaxy Survey (BGS), which contains >10 million galaxies across 14,000 deg² at z < 0.6. To characterize environment, we reconstruct the underlying large-scale density field using PyCosmoMMF, a NEXUS+-based algorithm that assigns each point in space to one of the four cosmic-web components: clusters, filaments, walls, or voids.

For morphology classification, we use a machine-learning classifier previously developed for the Siena Galaxy Atlas (SGA). The SGA is a catalog of bright, nearby galaxies constructed from the DESI Legacy Surveys. We adapt a pre-trained self-supervised algorithm, ssl-legacysurvey, that clusters galaxy images by color and morphology in a high-dimensional latent space and creates low-dimensional groupings of galaxies based on appearance. We then assign morphological types using a k-nearest-neighbor algorithm applied to the low-dimensional representation.

By combining these classifications with our density-field reconstruction, we quantify how morphology varies with the environment across a large volume of the low-redshift Universe. Although the morphology-density relation is well-established at z < 0.6, this work provides the foundation and methodology needed to extend the study to higher redshifts with future surveys.