Answers are grounded in this paper and cite their sources.
Next-generation galaxy surveys will map large-scale structure over unprecedented volumes, bringing two goals within reach: constraining primordial non-Gaussianity (PNG) and detecting relativistic effects on the largest scales. On these scales, relativistic and wide-separation (light-cone) effects share the scale dependence of local-type PNG, so they must be modelled to avoid biased inferences and to recover the true uncertainties. We present joint power spectrum and bispectrum forecasts that include these effects, with a multi-tracer bispectrum covariance including the odd-parity multipoles and leading non-Gaussian terms. Neglecting light-cone effects biases local-type fNL by ~2σ for a Euclid-like H$α$ survey and $>10σ$ for a MegaMapper-like LBG survey, mostly through lensing; wide-separation and local relativistic effects contribute $<1σ$. With these corrections included, marginalising over the evolution and magnification biases without informative priors increases the uncertainty on local-type fNL by ~60% for a Euclid-like power spectrum and by ~50% for a joint power spectrum and bispectrum analysis of the even multipoles. However, the odd-parity multipoles of the single-tracer bispectrum and the multi-tracer power spectrum, which are sensitive to leading-order relativistic corrections, break most of this degeneracy. Non-local shapes are less affected; for these, a bright-faint split of the MegaMapper-like survey improves the joint constraints on the equilateral and orthogonal shapes by a factor of ~2. Finally, local relativistic effects should be detectable at SNR ~13, with a bright-faint split of a DESI-like BGS survey, provided the magnification bias is known to a few per cent, but only marginally, SNR 2.5, for the higher-redshift surveys, and not with the single-tracer bispectrum dipole.
No takes yet. Share an insight, caveat, or question.
Addis et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: