AI-assisted superresolution cosmological simulations – II. Halo substructures, velocities, and higher order statistics

ABSTRACT In this work, we expand and test the capabilities of our recently developed superresolution (SR) model to generate high-resolution (HR) realizations full phase-space matter distribution, including both displacement velocity, from computationally cheap low-resolution (LR) cosmological N-body simulations. The SR enhances simulation resolution by generating 512 times more tracer particles, extending into deeply nonlinear regime where complex structure formation processes take place. We validate deploying in 10 simulations box size 100 h−1 Mpc, examine power spectra, bispectra, two-dimensional spectra redshift space. find generated field matches true HR result at per cent level down scales k ∼ h Mpc−1. also identify inspect dark haloes their substructures. Our generates visually authentic small-scale structures that cannot be resolved LR input, are good statistical agreement with real results. performs satisfactorily on halo occupation correlations space, pairwise velocity matching results comparable scatter, thus demonstrating its potential making mock catalogues. technique can a powerful promising tool for modelling galaxy physics large volumes.