Enhancing the localization of gravitational-wave hosts with SKYFAST: Rapid volume and inclination angle reconstruction

¹ Dipartimento di Fisica e Astronomia, Università degli Studi di Firenze, Via Sansone 1, Sesto Fiorentino, (Firenze) I-50019, Italy
² INFN, Sezione di Firenze, Sesto Fiorentino, (Firenze) I-50019, Italy
³ Dipartimento di Fisica “Enrico Fermi”, Università di Pisa, Largo Bruno Pontecorvo 3, Pisa I-56127, Italy
⁴ INFN, Sezione di Pisa, Largo Bruno Pontecorvo 3 Pisa I-56127, Italy
⁵ Dipartimento di Fisica e Astronomia “Galileo Galilei” Università di Padova, I-35131 Padova, Italy
⁶ INAF – Osservatorio Astronomico di Roma, Via Frascati 33, Monte Porzio Catone, (Rome) I-00078, Italy

^⋆ Corresponding authors; gabriele.demasi@unifi.it, giulia.capurri@df.unipi.it

Received: 22 July 2024
Accepted: 14 January 2025

Abstract

Context. The scientific impact of GW170817 strongly supports the suggestion that we need an efficient electromagnetic follow-up campaign for gravitational-wave event candidates. The success of these campaigns critically depends on a fast and accurate localization of the source.

Aims. We present SKYFAST, an algorithm for the rapid identification of gravitational-wave hosts to optimize electromagnetic follow-up searches. The goal is to produce a list of the galaxies within the localization volume, ranked by their probability of being the host, along with an estimate of the inclination angle conditioned on the position of each galaxy.

Methods.SKYFAST runs alongside a full parameter estimation (PE) algorithm, from which posterior samples are taken. These samples are then used to reconstruct an analytical posterior of the sky position, luminosity distance, and inclination angle using a Dirichlet process Gaussian mixture model, which is a nonparametric Bayesian method.

Results. We show that SKYFAST can reconstruct an accurate localization using only a fraction (∼10%) of the total posterior samples produced by the PE. Moreover, SKYFAST generates a ranked list of the most probable hosts from a galaxy catalog of choice in a few minutes. This list includes information on the inclination angle posterior conditioned on the position of each candidate host. This breaks the degeneracy between inclination angle and luminosity distance.

Conclusions. The reconstruction of the posterior using fewer samples than the full PE can lead to significant time savings, depending on the PE algorithm employed. This is crucial for identifying the electromagnetic counterpart. The inclusion of the inclination angle information conditioned on the position of each galaxy can lead to an optimized electromagnetic follow-up.