Constraining primordial and gravitational mode coupling with the position-dependent bispectrum of the large-scale structure

TL;DR

This paper introduces the position-dependent bispectrum as a new method to measure squeezed-limit trispectra in large-scale structure, forecasting its potential to constrain primordial non-Gaussianity parameters with future surveys.

Contribution

It generalizes the position-dependent power spectrum method to the bispectrum and trispectrum, providing forecasts for constraining primordial non-Gaussianity parameters.

Findings

Can constrain local-type g_NL to ~3 x 10^5 with large surveys

Including all bispectral configurations improves constraints

Method probes a broad family of trispectra beyond local g_NL

Abstract

We develop and study the position-dependent bispectrum. It is a generalization of the recently proposed position-dependent power spectrum method of measuring the squeezed-limit bispectrum. The position-dependent bispectrum can similarly be used to measure the squeezed-limit trispectrum in which one of the wavelengths is much longer than the other three. In this work, we will mainly consider the case in which the three smaller wavelengths are nearly the same (the equilateral configuration). We use the Fisher information matrix to forecast constraints on bias parameters and the amplitude of primordial trispectra from the position-dependent bispectrum method. We find that the method can constrain the local-type $g_{\rm NL}$ at a level of $\sigma(g_{\rm NL}^{\rm local}) \approx 3 \times 10^5$ for a large volume SPHEREx-like survey; improvements can be expected by including all the triangular configurations of the bispectra rather than just the equilateral configuration. However, the same measurement would also constrain a much larger family of trispectra than local $g_{\rm NL}$ model. We discuss the implications of the forecasted reach of future surveys in terms of super cosmic variance uncertainties from primordial non-Gaussianities.