Accelerating the 2-point and 3-point galaxy correlation functions using Fourier transforms

TL;DR

This paper demonstrates how Fourier Transform techniques can significantly speed up the computation of two-point and three-point galaxy correlation functions, especially useful for large upcoming survey datasets.

Contribution

It introduces FT-based methods for anisotropic 2PCF multipole moments and accelerates the 3PCF algorithm, reducing computational costs compared to traditional pair counting.

Findings

FT methods compute anisotropic 2PCF multipoles efficiently.

FT accelerates 3PCF calculations, reducing complexity.

Applicable to large datasets from upcoming surveys.

Abstract

Though Fourier Transforms (FTs) are a common technique for finding correlation functions, they are not typically used in computations of the anisotropy of the two-point correlation function (2PCF) about the line of sight in wide-angle surveys because the line-of-sight direction is not constant on the Cartesian grid. Here we show how FTs can be used to compute the multipole moments of the anisotropic 2PCF. We also show how FTs can be used to accelerate the 3PCF algorithm of Slepian & Eisenstein (2015). In both cases, these FT methods allow one to avoid the computational cost of pair counting, which scales as the square of the number density of objects in the survey. With the upcoming large datasets of DESI, Euclid, and LSST, FT techniques will therefore offer an important complement to simple pair or triplet counts.