REVIEW
Luminosity-weighted galaxy and gravitational wave cross-correlations
Cosmology and Nongalactic Astrophysics · Submitted 23 Oct 2026 · 56 references
Pith's verdict: A solid forecast paper: luminosity weighting is new to the Peak Sirens pipeline and the robustness result likely holds, but the headline H0 precision number is internally inconsistent and the central claim overstates what a one-parameter bias model actually tests.
T0 · AI review
Why Pith says this
Highlight any text to comment on it.
Referee report
The paper
The paper claims
This paper claims galaxy–GW cross-correlations recover the GW bias to 10% even when the true host galaxies are unseen, and keep $H_0$ and $\Omega_m$ unbiased at 0.5% and 5% under any host model.
The reading
What carries the argument
The load-bearing objects are the luminosity-weighted galaxy density contrast, the effective host bias $b_{\rm eff}(z)=\sum_L L^\beta b_L(z)/\sum_L L^\beta$, and the GW bias defined with respect to the luminosity-weighted galaxy auto-spectrum, $b_{\rm gw}\equiv C_\ell^{ggw}/C_\ell^g$. The pipeline samples binary-black-hole hosts from per-luminosity-bin galaxy maps with probability $p_{\rm host}(L)\propto L^\beta$, masks to a DESI DR3-like footprint, estimates masked-sky spectra with a quadratic estimator and Gaussian covariance, and approximates the theoretical cross-spectrum as the mean of 1000 simulations rescaled to other cosmologies and multiplied by the single free parameter $b_{\rm gw}$. The derived SNR scaling $\mathrm{SNR}^\ell_{ij}\propto (b_g^i)^2 b_{\rm gw}^j$ is what carries the argument that luminosity weighting can in principle boost the signal and that higher host bias sharpens the parameter forecasts.
Extended reading notes
Core claim
The central claim is that luminosity weighting in galaxy–GW cross-correlations is equivalent to re-weighting the effective linear bias of the GW host population, and that the inference is largely insensitive to which weighting one uses. Weighting galaxies by $L^\beta$ changes the effective bias to $b_{\rm eff}=\sum_L L^\beta b_L/\sum_L L^\beta$, and the signal-to-noise of each angular mode scales as $(b_g)^2 b_{\rm gw}$, where $b_{\rm gw}$ is defined relative to the luminosity-weighted galaxy auto-spectrum, $b_{\rm gw}\equiv C_\ell^{ggw}/C_\ell^g$. Because the five DESI luminosity bins have only mildly different biases, $\beta$ has a small effect on the contours; a toy model with a bright-bin bias four times larger shows the potential gain. The more consequential result is the host-mismatch test: when GW hosts are drawn from an unobserved galaxy population whose linear bias differs from the observed one, a template made of the fiducial cross-spectrum times a single free $b_{\rm gw}$ recovers the injected bias within 10% at $\ell_{\max}=110$. At $\ell_{\max}=190$, $b_{\rm gw}$ itself becomes biased, which the authors attribute to nonlinearities and the breakdown of scale-independent bias, yet $H_0$ and $\Omega_m$ remain unbiased and tighten to 0.5% and 5%.
Load-bearing premise
The 10% recovery claim for the GW bias assumes that a single scale-independent number $b_{\rm gw}$ multiplying the galaxy cross-spectrum describes the host population up to $\ell_{\max}=110$; the authors show that this assumption breaks at $\ell_{\max}=190$, where nonlinearities and scale-dependent bias contaminate $b_{\rm gw}$ even though the cosmological parameters stay unbiased.
Editorial extensions
If this is right
- If the robustness holds, future dark-siren cross-correlation analyses do not need a precise model of how galaxy luminosity selects GW hosts: the same cosmological answer emerges under uniform, linear, and quadratic weighting for DESI-like surveys.
- The GW bias becomes a measurable astrophysical quantity at about 10% accuracy with a third-generation network, encoding the clustering of the true (possibly unobserved) host population and, through it, compact-binary formation physics.
- The $H_0$ and $\Omega_m$ forecasts of 0.5% and 5% do not require a trustworthy $b_{\rm gw}$ model, so a stage-IV spectroscopic survey plus Einstein Telescope-class detectors can deliver a competitive, largely astrophysics-free Hubble-constant measurement.
- Restricting to large scales ($\ell_{\max}=110$) is the price of an unbiased $b_{\rm gw}$; the scale at which the linear-bias template fails could itself be used to probe the scale dependence of host galaxy bias.
Reading between the lines
- The insensitivity to luminosity weighting suggests the practical bottleneck for this method shifts from host-galaxy astrophysics to survey systematics such as completeness, photometric calibration, and mask errors, so tests on real data should prioritize those.
- The toy model with a bright-bin bias four times larger shows a regime where $\beta$ is measurable rather than a nuisance: a survey with a steeper luminosity–bias relation could turn the luminosity weight itself into a fitted astrophysical parameter probed by GW clustering.
- The same pipeline could test whether the 10% $b_{\rm gw}$ recovery survives a host population whose bias is scale-dependent below $\ell=110$, for example through assembly bias or a redshift-evolving merger rate; the paper's own $\ell_{\max}=190$ result shows where the linear template starts to bend.
- Comparing the recovered $b_{\rm gw}$ against host-galaxy constraints from galaxy–galaxy lensing or clustering would provide an independent consistency check of the linear-bias assumption in real data.
Original abstract
Cross-correlations of galaxies and gravitational waves (GW) can be used to extract cosmological parameters and the gravitational wave bias. In this work, we study how to include galaxy luminosities in these spatial cross-correlations, inspired by the standard approach of dark sirens line-of-sight redshift priors. We present a formalism to include luminosity information in galaxy-GW simulations and study the effects of assigning a hosting probability that depends on luminosity through a simple power law $p_{\rm{host}} \propto L^\beta$ with $\beta \in \{0,1,2\}$ corresponding to uniformly, linear and quadratically weighted in luminosity. This effectively changes the GW linear bias with respect to dark matter and, in principle, allows to gain signal-to-noise and a sharper measurement of cosmological parameters. For a realistic luminosity dependence of the galaxy bias taken from DESI data, the gain is marginal, making the method robust to different luminosity weighting prescriptions. We show that the cross-correlation inference is able to recover the injected GW bias within $10\%$ if there is a mismatch between the injected linear bias of host galaxies and observed galaxies, provided that one sticks to relatively large scales. We also show that the Hubble constant and matter energy density can be recovered to $0.5\%$ and $5\%$ respectively with a third generation network of GW detectors, regardless of the modeling of the GW bias.
The compiled record
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
The abstract, sentence by sentence
Cross-correlations of galaxies and gravitational waves (GW) can be used to extract cosmological parameters and the gravitational wave bias. In this work, we study how to include galaxy luminosities in these spatial cross-correlations, inspired by the standard approach of dark sirens line-of-sight redshift priors. We present a formalism to include luminosity information in galaxy-GW simulations and study the effects of assigning a hosting probability that depends on luminosity through a simple power law $p_{\rm{host}} \propto L^\beta$ with $\beta \in \{0,1,2\}$ corresponding to uniformly, linear and quadratically weighted in luminosity. This effectively changes the GW linear bias with respect to dark matter and, in principle, allows to gain signal-to-noise and a sharper measurement of cosmological parameters. For a realistic luminosity dependence of the galaxy bias taken from DESI data, the gain is marginal, making the method robust to different luminosity weighting prescriptions. We show that the cross-correlation inference is able to recover the injected GW bias within $10\%$ if there is a mismatch between the injected linear bias of host galaxies and observed galaxies, provided that one sticks to relatively large scales. We also show that the Hubble constant and matter energy density can be recovered to $0.5\%$ and $5\%$ respectively with a third generation network of GW detectors, regardless of the modeling of the GW bias.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Cross-correlations of galaxies and gravitational waves (GW) can be used to extract cosmological parameters and the gravitational wave bias. | background 1.00 | 0.07 | 0.05 | asserted 0.51 |
| s1 | In this work, we study how to include galaxy luminosities in these spatial cross-correlations, inspired by the standard approach of dark sirens line-of-sight redshift priors. | contribution 0.97 | 0.69 | 0.03 | asserted 0.89 |
| s2 | We present a formalism to include luminosity information in galaxy-GW simulations and study the effects of assigning a hosting probability that depends on luminosity through a simple power law $p_{\rm{host}} \propto L^\beta$ with $\beta \in \{0,1,2\}$ corresponding to uniformly, linear and quadratically weighted in luminosity. | method 0.74 | 0.94 | 0.07 | asserted 0.99 |
| s3 | This effectively changes the GW linear bias with respect to dark matter and, in principle, allows to gain signal-to-noise and a sharper measurement of cosmological parameters. | interpretation 0.97 | 0.31 | 0.20 | hedged possibility 0.85 |
| s4 | For a realistic luminosity dependence of the galaxy bias taken from DESI data, the gain is marginal, making the method robust to different luminosity weighting prescriptions. | result 0.86 | 0.89 | 0.39 | asserted 0.99 |
| s5 | We show that the cross-correlation inference is able to recover the injected GW bias within $10\%$ if there is a mismatch between the injected linear bias of host galaxies and observed galaxies, provided that one sticks to relatively large scales. | result 0.99 | 0.95 | 0.12 | asserted 0.99 |
| s6 | We also show that the Hubble constant and matter energy density can be recovered to $0.5\%$ and $5\%$ respectively with a third generation network of GW detectors, regardless of the modeling of the GW bias. | result 1.00 | 0.94 | 0.33 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
Figures
Showing 8 of the paper's 13 figures. See all figures in the PDF ↗ Figures reproduced from arXiv:2610.32436 by the authors.
References
Discussion
Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.