Accurate emulator for the redshift-space power spectrum of dark matter halos and its application to galaxy power spectrum
An accurate theoretical template of the redshift-space galaxy power spectrum, if applicable out to nonlinear scales, enables us to extract more stringent and robust constraints on cosmological parameters from the measured galaxy clustering. In this work, we develop a simulation-based template, so-called emulator, for the redshift-space power spectrum of dark matter halos. Using the redshift-space halo power spectra measured from the DARK QUEST N -body simulation suite that covers 101 flat-geometry w -cold dark matter (w CDM ) cosmologies around the Planck Λ CDM model, we feed these data into a feed-forward neural network to build the fast and accurate emulation of the power spectrum from the linear to nonlinear scales up to k ≃0.6 h Mpc-1 . Our emulator achieves about 1% and 5% fractional accuracies in predicting the monopole and quadrupole moments of the power spectrum, respectively, for halos of ∼1013h-1 M⊙ that correspond to host halos of the Sloan Digital Sky Survey (SDSS) LOWZ- and CMASS (constant mass)-like galaxies, where the achieved accuracies are sufficient compared to the statistical errors of SDSS volume. The validation and performance of the emulator are given by the comparison of the emulator predictions with the power spectra directly measured from the simulations for validation sets that are not used in the training. We demonstrate that the emulator outputs can be used to make model predictions for the redshift-space power spectrum of galaxies by employing user-fed models for the halo-galaxy connection, such as the halo occupation distribution. The emulator allows us to easily incorporate the Finger-of-God effect due to the virial motions of galaxies and the Alcock-Paczyński distortions. Our code can compute the redshift-space galaxy power spectrum in a CPU subseconds and is ready to perform the emulator-based cosmological analysis for the exiting and upcoming galaxy redshift surveys.