Presentation #326.01 in the session Gravitational Wave Cosmology and Methodologies.
The Hubble tension between early-time and late-time measurements of the expansion rate of the universe now stands at a 5-σ disparity. Another method of measuring the Hubble constant that is independent of the two current methods is needed to determine if new physics are necessary to explain the discrepancy. One promising approach is to use gravitational wave standard sirens, in which a binary neutron star (BNS) merger’s redshift is inferred through its electromagnetic (EM) emission while its luminosity distance is measured from its gravitational wave (GW) emission, enabling an inference on the Hubble constant via Hubble’s law. However, this approach suffers from systematic biases arising from the geometry of the BNS mergers themselves. In rough terms, there are two biases: that arising from the anisotropy of the GW emission, and that arising from the anisotropy of the EM emission. In each bias, an event viewed on the angular momentum axis is more likely to be detected than an oblique event, pushing the inferred luminosity distance nearer to the observer, and thus inflating the inferred Hubble constant. Furthermore, there are two classes of BNS merger: those with detected EM counterparts (standard sirens), and those without. Previous attempts at bias removal have considered only one of the biases independently of the other and have only considered data from a suite of standard sirens. We present a novel method of simulation-based inference which leverages data from both event classes to correct for both biases at once.