Dark Siren Cosmology with the Dark Energy Spectroscopic Instrument (DESI)

We present new results from a joint cosmological analysis combining the latest gravitational wave catalog, GWTC-4, with proprietary data from the Dark Energy Spectroscopic Instrument (DESI). Our work focuses on dark siren cosmology, leveraging gravitational wave (GW) events without uniquely identified electromagnetic counterparts to statistically infer cosmological parameters. DESI's unprecedented spectroscopic mapping of the large-scale structure of the Universe enables precise modeling of the galaxy distribution within GW localization volumes. This is particularly powerful for low-redshift (z<0.4) events, where DESI's completeness allows robust application of the dark siren framework. However, GWTC-4 includes events extending to z∼1, where spectroscopic coverage is sparse. To address this, we consistently employ the spectral siren framework, jointly inferring the black hole mass distribution and cosmological parameters within our dark siren analyses. A key innovation in our analysis is the use of Gaussian Process models to describe the GW source population. This flexible, data-driven approach captures features in the mass spectrum—that are missed by traditional parametric models. Neglecting such features can bias cosmological inference. Our inference codes are written in JAX, allowing us to fully exploit GPU acceleration to handle the computational demands of analyzing the GWTC-4 data along with millions of DESI galaxies. This work sets the stage for future joint analyses with upcoming DESI data releases and next-generation GW catalogs.