Efficient computation of the galaxy angular bispectrum in redshift space

Anthony Challinor · 2026

Read Efficient computation of the galaxy angular bispectrum in redshift space by Anthony Challinor (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-30246.txt.

Introduction

Zucheng Gao, Zvonimir Vlah and Anthony Challinor compute the tree-level unequal-time bispectrum in real and redshift space. Results express the bispectrum as a sum of products of two power spectra, enabling a flat-sky approximation that improves efficiency. It preserves line-of-sight information while discarding extraneous full-sky contributions. Validation against full-sky integration shows agreement across scales and redshifts, and at redshift one, sub-percent agreement for multipoles over five. At small scales, results align with the Limber approximation, and Python code is on GitHub.

Catalog record

CREATORAnthony Challinor
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE82968
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.30246

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Efficient computation of the galaxy angular bispectrum in redshift space Anthony Challinor Abstract Efficient computation of the angular bispectrum is an essential part of modelling large-scale structure observations, but it still remains an extremely challenging task. In this work, we compute the tree-level, unequal-time angular bispectrum in both real and redshift space. By deriving full-sky results, we show that the bispectrum can be expressed as a sum of products of two angular power spectra, enabling the use of our recently developed flat-sky approximation to enhance computational efficiency significantly. This flat-sky formalism preserves key line-of-sight mode information while discarding extraneous full-sky contributions. We validate our approach by comparing it with direct full-sky integration, finding excellent agreement across a wide range of scales and redshifts for all bispectrum configurations. At redshift z = 1 z=1 , we achieve sub-percent agreement (for multipoles ℓ ≳ 5 \ell\gtrsim 5 ) between full-sky and flat-sky results for equilateral, squeezed, and folded configurations, using narrow Gaussian radial window functions ( σ z = 0.01 \sigma_{z}=0.01 ) in both equal-time and unequal-time scenarios. On small scales, where direct full-sky integration becomes computationally prohibitive, our results align with the Limber approximation (where applicable), confirming the robustness and accuracy of our implementation. To facilitate future studies, we provide a Python implementation of our results, which is publicly available on GitHub . 1 Introduction With powerful new galaxy surveys with instruments such as Euclid [ 21 ] and DESI [ 8 ] underway, and the imminent Rubin Observatory [ 22 ] and launch of the Roman Telescope [ 26 ] , we have entered a new era of precision cosmology. These surveys have large sky coverage and high survey depth, resulting in a high number of detected galaxies. To extract cosmological information reliably from these large datasets, the theoretical modeling needs to be highly accurate and computationally efficient. The three-point correlation function, or bispectrum (its Fourier counterpart), is an important quantity to study primordial non-Gaussianity and other potential additions to the standard \Lambda CDM model, such as modified gravity. Even within the standard model, the bispectrum can play an important role in breaking degeneracies between the growth of structure and galaxy bias. Including redshift space distortions (RSD) is essential in modelling the bispectrum since their effect can partly mimic other large-scale signals of interest, such as local-model primordial non-Gaussianity. The 3D bispectrum (in Cartesian Fourier coordinates) with RSD included has been investigated intensively in the literature [ 15 , 16 , 7 , 19 , 9 ] . These calculations usually assume the plane-parallel (flat-sky) approximation, which makes the subsequent addition of wide-angle effects a daunting task [ 24 , 5 ] . Alternatively, the full-sky angular bispectrum (which is directly observable) intrinsically includes these projection effects, and can be naturally extended to include the full set of relativistic effects [ 28 , 11 , 23 ] . However, the direct computation of the angular bispectrum is a computationally challenging task, so it is seldom used in practical data analysis. Even in its simplest form, considering only second-order biasing, modelling the projected full-sky angular bispectrum is computationally demanding. To address this, reference [ 3 ] introduced an FFTLog-based approach that leverages the separability of the tree-level bispectrum to accelerate computations. Despite this improvement, the method still depends on complex hypergeometric functions, which are expensive to evaluate. Subsequently, reference [ 10 ] extended this work to redshift space, while reference [ 23 ] incorporated a set of the relativistic effects. In practice, additional integration of survey-specific radial window functions is needed to obtain the observed angular bispectrum, making this process too slow for a practical Markov Chain Monte Carlo (MCMC) analysis. To overcome these difficulties, one may turn to machine learning interpolation and emulation techniques, though these inevitably come with additional drawbacks, like lack of physical insights and lead to a lack of interpretability. We focus on alternative strategies based on analytical and computational simplifications. In our previous work [ 14 ] , we showed that the flat-sky result (derived as a leading contribution in the asymptotic expansion) can approximate the full-sky angular power spectrum (two-point correlation function) to a percent-level even on the largest scales. At the same time, it can be computed numerically very efficiently, making it especially suitable for the MCMC analysis. In this work, we apply our flat-sky approximation to the angular bispectrum. We first (re)derive the general form for the tree-level full-sky angular bispectrum with RSD, starting from the observed galaxy number counts in redshift space at second order in perturbation theory, which also includes wide-angle effects and non-distant-observer corrections. We start directly from the full-sky angular bispectrum and approximate integrals over wavenumber, which involve products of spherical Bessel functions, with their flat-sky Fourier equivalents by generalising results from Gao et al. [ 13 ] . This method naturally results in a set of correction terms, which improve the accuracy in approximating the full-sky angular bispectrum. In other words, by using the correspondence between full- and flat-sky integrals on the general result for the full-sky angular bispectrum, we naturally obtain correction terms for wide-angle effects. The paper is organised as follows: in section 2 , we first present our real-space results for the unequal-time angular bispectrum. We extend these to redshift space in section 3 . We conclude in section 4 . In appendix A , we check the validity of the distant-observer approximation. We work with the Planck , flat, \Lambda\text{CDM} cosmology [ 1 ] , with CDM density \Omega_{c}h^{2}=0.11933 , baryon density \Omega_{b}h^{2}=0.02242 , Hubble constant H_{0}=100h\,\text{km\,s}^{-1}\,\text{Mpc}^{-1} with h=0.6766 , scalar spectral index n_{s}=0.9665 , and fluctuation amplitude \sigma_{8}=0.8103 . All results presented here were obtained using the Python code that we made publicly available on GitHub . 1 1 1 https://github.com/GZCPhysics/Bispectrum.git 2 Angular bispectrum in real space In this section, we focus on the real-space angular bispectrum of a biased tracer at leading order (tree-level) in cosmological perturbation theory. We outline the derivation of the full-sky angular bispectrum and detail the numerical optimisation techniques employed in its computation. Additionally, we compare with the Limber approximation (while we discuss an extended version specifically for the squeezed limit in appendix B ). Finally, we present numerical results for the real-space bispectrum, highlighting key features and implications. 2.1 Real space overdensity and the angular bispectrum To obtain the tree-level bispectrum results, let us first consider the expansion of the overdensity field of a biased tracer up to the second-order: \displaystyle\delta(\bm{x};\;z)=\delta^{(1)}(\bm{x};\;z)+\delta^{(2)}(\bm{x};\;z)+\cdots\,, (2.1) where each order can be related to the linear matter overdensity field, \delta_{\rm G} (the subscript “G” indicates that the field is Gaussian). In Fourier space, we have \displaystyle\delta^{(1)}(\bm{k};\;z) \displaystyle=b_{10}(z)\delta_{\mathrm{G}}(\bm{k};\;z)\,, (2.2) \displaystyle\delta^{(2)}(\bm{k};\;z) \displaystyle=\int\frac{d^{3}\bm{k}_{1}}{(2\pi)^{3}}\frac{d^{3}\bm{k}_{2}}{(2\pi)^{3}}(2\pi)^{3}\delta^{\rm D}(\bm{k}-\bm{k}_{1}-\bm{k}_{2})\mathcal{K}_{2}(\bm{k}_{1},\bm{k}_{2};\,z)\delta_{\mathrm{G}}(\bm{k}_{1};\;z)\delta_{\mathrm{G}}(\bm{k}_{2};\;z)\,, (2.3) where we introduced the second-order tracer kernel \mathcal{K}_{2}(\bm{k}_{1},\bm{k}_{2};\,z)=b_{10}(z)\mathcal{F}_{2}(\bm{k}_{1},\bm{k}_{2})+b_{20}(z)+b_{K^{2}}(z)\mathcal{S}_{2}(\bm{k}_{1},\bm{k}_{2})\,. (2.4) Alongside the quadratic bias term ( b_{20} ), this contains the second-order matter kernel \mathcal{F}_{2} and shear kernel \mathcal{S}_{2} , explicitly given by \displaystyle\mathcal{F}_{2}(\bm{k}_{1},\bm{k}_{2}) \displaystyle=\frac{5}{7}+\frac{1}{2}\frac{\bm{k}_{1}\cdot\bm{k}_{2}}{k_{1}k_{2}}\left(\frac{k_{1}}{k_{2}}+\frac{k_{2}}{k_{1}}\right)+\frac{2}{7}\left(\frac{\bm{k}_{1}\cdot\bm{k}_{2}}{k_{1}k_{2}}\right)^{2} \displaystyle=\frac{5}{14}-\frac{5}{28}\frac{k_{1}^{2}}{k_{2}^{2}}-\frac{5}{28}\frac{k_{2}^{2}}{k_{1}^{2}}+\frac{3}{28}\frac{k^{2}}{k_{1}^{2}}+\frac{3}{28}\frac{k^{2}}{k_{2}^{2}}+\frac{1}{14}\frac{k^{4}}{k_{1}^{2}k_{2}^{2}}\,, (2.5) \displaystyle\mathcal{S}_{2}(\bm{k}_{1},\bm{k}_{2}) \displaystyle=\frac{(\bm{k}_{1}\cdot\bm{k}_{2})^{2}}{k_{1}^{2}k_{2}^{2}}-\frac{1}{3} \displaystyle=\frac{1}{6}+\frac{1}{4}\frac{k_{1}^{2}}{k_{2}^{2}}+\frac{1}{4}\frac{k_{2}^{2}}{k_{1}^{2}}-\frac{1}{2}\frac{k^{2}}{k_{2}^{2}}-\frac{1}{2}\frac{k^{2}}{k_{1}^{2}}+\frac{1}{4}\frac{k^{4}}{k_{1}^{2}k_{2}^{2}}\,, (2.6) where we have implicitly used the triangle relation, \bm{k}=\bm{k}_{1}+\bm{k}_{2} , to write the kernels in a separable form. Note that we have made the quasi-EdS approximation for the time-dependence of the second-order matter density in writing down a time-independent matter kernel \mathcal{F}_{2} . In addition to the deterministic bias contributions,

Read the complete text