Abstract: Local primordial non-Gaussianity (LPNG) is predicted by many non-minimal models of inflation, and creates a scale-dependent contribution to the power spectrum of large-scale structure (LSS) tracers, whose amplitude is characterized by bϕ . Knowledge of bϕ for the observed tracer population is therefore crucial for learning about inflation from LSS. Recently, it has been shown that the relationship between linear bias b1 and bϕ for simulated halos exhibits significant secondary dependence on halo concentration. We leverage this fact to forecast multi-tracer constraints on f loc NL. We train a machine learning model on observable properties of simulated IllustrisTNG galaxies to predict bϕ for samples constructed to approximate DESI emission line galaxies (ELGs) and luminous red galaxies (LRGs). We find σ(f loc NL) = 2.3, and σ(f loc NL = 3.7, respectively. These forecasted errors are roughly factors of 3, and 35% improvements over the single-tracer case for each sample, respectively. When considering both ELGs and LRGs in their overlap region, we forecast σ(f loc NL) = 1.5 is attainable with our learned model, more than a factor of 3 improvement over the single-tracer case, while the ideal split by bϕ could reach σ(f loc NL) < 1. We also perform multi-tracer forecasts for upcoming spectroscopic surveys targeting LPNG (MegaMapper, SPHEREx) and show that splitting tracer samples by bϕ can lead to an order-of-magnitude reduction in projected σ(f loc NL for these surveys.

We update Halo Zeldovich Perturbation Theory (HZPT, [1]), an analytic model for the two-point statistics of dark matter, to describe halo and galaxy clustering, and galaxy-matter cross-correlation on nonlinear scales. The model correcting Zeldovich has an analytic Fourier transform, and therefore is valid in both configuration space and Fourier space. The model is accurate at the 2%-level or less for Pmm (k < 1 h/Mpc), Phm (k < 1 h/Mpc), Phh (k < 2 h/Mpc), Pgm (k < 1 h/Mpc), Pgg (k < 1 h/Mpc), ξmm (r > 1 Mpc/h), ξhm (r > 2 Mpc/h), ξhh (r > 2 Mpc/h), ξgm (r > 1 Mpc/h), ξgg (r > 2 Mpc/h), for LRG-like mock galaxies. We show that the HZPT model for matter correlators can account for the effects of a wide range of baryonic feedback models and provide two extended dark matter models which are of 1% (3%) accuracy for k < 10 (8) h/Mpc. We explicitly model the non-perturbative features of halo exclusion for the halo-halo and galaxy-galaxy correlators, as well as the presence of satellites for galaxy-matter and galaxy-galaxy correlation functions. We perform density estimation using N-body simulations and a wide range of HOD galaxy mocks to obtain correlations of model parameters with the cosmological parameters Ωm and σ8. HZPT can provide a fast, interpretable, and analytic model for combined-probe analyses of redshift surveys using scales well into the non-linear regime.