Abstract Third-generation (3G) gravitational-wave (GW) detectors will enable precision cosmology with dark and bright sirens, motivating realistic simulations that jointly model GW sources, electromagnetic (EM) counterparts, and host galaxies. We present a self-consistent framework that embeds binary compact object mergers within galaxies from the CosmoDC2 LSST synthetic catalog. We model ab initio the mass, spin, and redshift distributions of binary black holes (BBH), binary neutron stars (BNS), and black hole-neutron star (BHNS) systems using astrophysically motivated prescriptions, and sample host galaxies using redshift- and mass-dependent merger-rate templates calibrated on population-synthesis results. We generate GW signals for multiple waveform approximants and detector networks spanning second-generation (LIGO, Virgo, KAGRA) and third-generation observatories (Einstein Telescope, Cosmic Explorer), including realistic duty cycles. Parameter covariances are estimated with Fisher-matrix methods, while EM counterparts are modeled via kilonova prescriptions to produce synthetic LSST-band photometry. We publicly release CosmoDC2BCO, providing intrinsic and extrinsic source parameters, signal-to-noise ratios, parameter uncertainties, sky areas, and kilonova magnitudes. We find that 3G networks dramatically increase detection rates and improve parameter estimation, and that retaining 2G detectors alongside 3G facilities can significantly enhance sky localization and distance precision, particularly for BNS. Under a simplified Target-of-Opportunity strategy, an LSST-like survey paired with a CE+ET+LVK network at 70% duty cycle could detect about 5 000 GW-associated kilonovae over 10 years on a 16 000 deg2 footprint. These forecasts depend on merger-rate and kilonova-luminosity assumptions and should be interpreted accordingly.
Menote et al. (2026) studied this question.