Galaxy bispectrum, primordial non-Gaussianity and redshift space distortions (1603.06814v2)

Abstract: Measurements of the non-Gaussianity of the primordial density field have the power to considerably improve our understanding of the physics of inflation. Indeed, if we can increase the precision of current measurements by an order of magnitude, a null-detection would rule out many classes of scenarios for generating primordial fluctuations. Large-scale galaxy redshift surveys represent experiments that hold the promise to realise this goal. Thus, we model the galaxy bispectrum and forecast the accuracy with which it will probe the parameter $f_{\rm NL}$, which represents the degree of primordial local-type non Gaussianity. Specifically, we address the problem of modelling redshift space distortions (RSD) in the tree-level galaxy bispectrum including $f_{\rm NL}$. We find novel contributions associated with RSD, with the characteristic large scale amplification induced by local-type non-Gaussianity. These RSD effects must be properly accounted for in order to obtain un-biased measurements of $f_{\rm NL}$ from the galaxy bispectrum. We propose an analytic template for the monopole which can be used to fit against data on large scales, extending models used in the recent measurements. Finally, we perform idealised forecasts on $\sigma_{f_{\rm NL}}$ -- the accuracy of the determination of local non-linear parameter $f_{\rm NL}$ -- from measurements of the galaxy bispectrum. Our findings suggest that current surveys can in principle provide $f_{\rm NL}$ constraints competitive with Planck, and future surveys could improve them further.