First measurement of the Hubble constant from gravitational wave-galaxy cross-correlations

Isabela Santiago de Matos, Charles Dalang, Tessa Baker, Raul Abramo, João Ferri, Miguel Quartin · 2025

Read First measurement of the Hubble constant from gravitational wave-galaxy cross-correlations by Isabela Santiago de Matos, Charles Dalang, Tessa Baker, Raul Abramo, João Ferri, Miguel Quartin (2025) 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-2512-15380.txt.

Introduction

Isabela Santiago de Matos and five coauthors measure the expansion rate of the universe with a brand new cross check. Their method, Peak Sirens, cross correlates galaxy positions with gravitational wave distances. The peak of that cross spectrum, over redshift and luminosity distance, reveals the Hubble constant.

Using two gravitational wave events from the GWTC-3.0 catalog and the GLADE+ galaxy catalog, they detect the cross correlation peak for the first time, at 5.9 sigma confidence. Most of the signal comes from GW190814, the best localized event, at 3.4 sigma alone. Adding the famous GW170817 event, but deliberately ignoring its known redshift, gives a Hubble constant of 67. The errors run 18 up and 15 down, in kilometers per second per megaparsec. They also bound the gravitational wave bias below 4.3 at 95 percent confidence.

Catalog record

CREATORIsabela Santiago de Matos, Charles Dalang, Tessa Baker, Raul Abramo, João Ferri, Miguel Quartin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-15380

Browse the full library · About the Tower of Babel

Full text

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

First measurement of the Hubble constant from gravitational wave-galaxy cross-correlations Preprint: BLA Isabela Santiago de Matos, Charles Dalang, Tessa Baker, Raul Abramo, João Ferri, Miguel Quartin Abstract We measure for the first time the Hubble constant ( H 0 H_{0} ) from the cross-correlation of galaxies and gravitational waves (GW), by applying the Peak Sirens method. This method consists of finding the peak of the 3D angular cross-spectrum C ℓ ​ ( z , D L ) C_{\ell}(z,D_{L}) between the galaxy redshifts ( z z ) and the GW luminosity distances ( D L D_{L} ). Using two GW events from the GWTC-3.0 catalog and the GLADE+ galaxy catalog, we make the first detection of the cross-correlation peak at 5.9 ​ σ 5.9\sigma confidence. This signal comes mostly from the best localized event in the catalog, GW190814, which alone provides a 3.4 ​ σ 3.4\sigma significance. Adding also the multimessenger event GW170817, but without using its known redshift, we find H 0 = 67 − 15 + 18 H_{0}=67^{+18}_{-15} km s -1 Mpc -1 and the first observational constraint on the GW bias, b gw < 4.3 b_{\rm gw}<4.3 at 95% CI. These measurements set the stage for future novel cosmological constraints with this technique. Introduction. The past decade of gravitational wave (GW) observations has opened previously uncharted regimes of gravity, cosmology, and astrophysics 67 ; 30 . For the new field of GW cosmology, the primary focus to date has been constraints on the luminosity distance-redshift relation, which is sensitive to both cosmological parameters and departures from General Relativity (GR) on cosmological scales. Such constraints are possible because gravitational waveforms from compact binary coalescences (CBCs) inherently yield the luminosity distance D_{L} of the source (but not its redshift z ); this is in contrast with most electromagnetic (EM) sources, whose redshifts, unlike their distances, are generally straightforward to measure. Several techniques have been proposed to recover the redshifts of GW sources, with most effort focused on bright siren 92 ; 53 ; 34 ; 78 ; 77 ; 17 ; 69 and dark siren methods 92 ; 64 ; 38 ; 66 ; 28 . At present, the multimessenger event GW170817 remains the only confirmed bright siren 7 ; 6 . Consequently, the dark siren method has risen to become the current workhorse of GW cosmology – see applications to real data in 45 ; 93 ; 81 ; 49 ; 87 ; 44 ; 82 ; 51 ; 9 ; 10 ; 57 ; 32 ; 27 ; 2 . In particular, the most recent bound on the Hubble constant ( H_{0} ) obtained by LIGO-Virgo-KAGRA (LVK) 5 ; 1 ; 12 ; 15 ; 20 ; 94 from a combined dark+bright siren analysis of GWTC-4.0 3 ; 31 ; 95 together with the GLADE+ galaxy catalog 35 is: H_{0}=76.6^{+13.0}_{-9.5} km s -1 Mpc -1 at 68.3\% confidence interval (CI) 2 . In this work we apply to current data a third, different technique for associating redshift information to GW sources: GW-galaxy cross-correlation. Several variants of this technique exist 73 ; 71 ; 40 ; 46 ; 22 ; 100 ; 13 ; 85 ; 88 ; 83 , but in particular we adopt the Peak Sirens method put forward in 43 , which draws upon on the original work of 79 , an idea akin to the photometric redshift calibration of 74 . The Peak Sirens method uses the fact that the spatial correlations between GW events and galaxies, measured in terms of the angular cross-spectrum, peak when the redshifts of galaxies correspond to the luminosity distances of the GW hosts. Like other standard siren methods, Peak Sirens bypasses the need for external distance calibrations, making it a promising avenue for tackling the Hubble tension 39 ; the position of the peak was also shown in 43 to be a robust, model-independent cosmic standard ruler which is insensitive to the precise shape of the matter power spectrum. The Peak Sirens method allows constraints on the bias of GW sources ( b_{\rm gw} ), a number quantifying the extent to which GW sources trace the underlying dark matter field. This is a point of difference with current dark sirens approaches, which do not have direct sensitivity to b_{\rm gw} . Another crucial point of difference with the dark sirens method is that the Peak Sirens technique does not rely on assumptions about the mass distribution and merger rate of CBCs. On one hand, features in the mass distribution are a highly informative component of current constraints, especially when the galaxy catalog used has low completeness over the typical range of the GW events in hand 41 ; 86 ; 58 ; 42 ; 65 . On the other hand, this dependence introduces a large number of additional parameters that must be constrained alongside the cosmology – up to 25 in 2 . Hence, Peak Sirens is a more direct probe of cosmology with less astrophysical entanglement. In this letter we perform the first application of the Peak Sirens method to data using GW190814 and GW170817 (though the redshift of its host galaxy is not used) and the K-band galaxies of the GLADE+ catalog (see however 1 1 1 We note that an attempt to detect the GW-galaxy cross-correlation signal using similar data sets was presented in 70 , but returning a null result. This work has various differences in methodology compared to ours, which do not allow it to precisely locate the peak of GW-galaxy correlations computed in pairs of thin and widely separated radial shells. ). As very few events lie within the depth of the GLADE+ catalog, GW170817 is included here to help demonstrate the viability of the method, even if tighter H_{0} constraints might be obtained instead with its use as a bright siren. Since we have good galaxy coverage for only a limited number of GWs, we focus on the dominant parameter sensitivities: H_{0} and b_{\rm gw} . Methodology 3D angular cross-correlations. A key assumption of this work is that GW sources are located inside galaxies, such that their positions are correlated. To study these correlations, we build pixelated galaxy and GW source sky maps using HEALPix 101 ; 47 . For galaxies, we decompose the radial direction in redshift bins z_{i} , which, together with the HEALPix pixels, fully defines voxels. We count the number of galaxies N_{\mathrm{g}}(z_{i},\boldsymbol{\hat{n}}) in each voxel denoted by the direction \boldsymbol{\hat{n}} and redshift bin center z_{i} . We integrate the z -PDF assuming Gaussian uncertainties over each voxel. This map encodes the PDF of finding galaxies in each voxel. For the GW data, we decompose instead the radial direction in D_{L} bins indicated by their center D_{Lj} . We integrate the observed 3D location posterior, marginalized over other binary parameters like masses and spins, to count the fractional GW contribution to each voxel. This defines the GW map N_{\mathrm{gw}}(D_{Lj},\boldsymbol{\hat{n}}) , which is proportional to the PDF of GW sources. With these two maps, one can compute their angular cross-correlation: \displaystyle\textstyle C_{\ell}(z_{i},D_{Lj})=\frac{1}{2\ell+1}\sum_{m=-\ell}^{\ell}a_{\ell m}^{\mathrm{g}}(z_{i})\,a^{\mathrm{gw}*}_{\ell m}(D_{Lj})\,, (1) where a_{\ell m}^{\mathrm{g}}(z_{i}) and a_{\ell m}^{\mathrm{gw}}(D_{Lj}) are the coefficients of their respective spherical harmonic decompositions: \textstyle N_{X}(r_{Xi},{\boldsymbol{\hat{n}}})=\sum_{\ell m}a^{i}_{\ell m}(r_{Xi})Y_{\ell m}({\boldsymbol{\hat{n}}})\,, (2) where in turn

Read the complete text