Galaxy clustering from the bottom up: A Streaming Model emulator I

ABSTRACT In this series of papers, we present a simulation-based model for the non-linear clustering galaxies based on separate modelling in real space and velocity statistics. first paper, an emulator real-space correlation function galaxies, whereas real-to-redshift mapping statistics is presented second paper. Here, show that neural network galaxy trained data extracted from dark quest suite N-body simulations achieves sub-per cent accuracies scales 1 < r 30 $h^{-1} \, \mathrm{Mpc}$, better than 3 per $h^{-1}\, \mathrm{Mpc}$ predicting dark-matter haloes with number density 10−3.5$(h^{-1}\, \mathrm{Mpc})^{-3}$, close to SDSS LOWZ-like galaxies. The halo can be combined galaxy–halo connection predict through model. We demonstrate accurately recover cosmological parameters when depends only mass galaxies’ host halos. Furthermore, constraining power σ8 increases by about factor 2 including smaller 5 \mathrm{Mpc}$. However, not property responsible clustering, as observed hydrodynamical or semi-analytic models formation, our gives biased constraints σ8. This bias disappears small (r 10 \mathrm{Mpc}$) are excluded analysis. shows vanilla could introduce biases into analysis future sets.