Fast Parameter Estimation of Binary Mergers for Multimessenger Followup

TL;DR

This paper introduces a rapid Bayesian parameter estimation method for gravitational-wave events, enabling quick source characterization within 20 minutes to optimize electromagnetic followup efforts.

Contribution

It extends the relative binning technique to a coherent detector-network statistic for fast, unbiased parameter estimation from gravitational-wave data.

Findings

Produces marginalized posterior densities within 20 minutes

Achieves accuracy comparable to standard likelihood methods

Applicable to real-time gravitational-wave event analysis

Abstract

Significant human and observational resources have been dedicated to electromagnetic followup of gravitational-wave events detected by Advanced LIGO and Virgo. As the sensitivity of LIGO and Virgo improves, the rate of sources detected will increase. Margalit & Metzger (2019; arXiv:1904.11995) have suggested that it may be necessary to prioritize observations of future events. Optimal prioritization requires a rapid measurement of a gravitational-wave event's masses and spins, as these can determine the nature of any electromagnetic emission. We extend the relative binning method of Zackay et al. (2018; arXiv:1806.08792) to a coherent detector-network statistic. We show that the method can be seeded from the output of a matched-filter search and used in a Bayesian parameter measurement framework to produce marginalized posterior probability densities for the source's parameters within 20 minutes of detection on 32 CPU cores. We demonstrate that this algorithm produces unbiased estimates of the parameters with the same accuracy as running parameter estimation using the standard gravitational-wave likelihood. We encourage the adoption of this method in future LIGO-Virgo observing runs to allow fast dissemination of the parameters of detected events so that the observing community can make best use of its resources.