Pith. sign in

REVIEW 3 major objections 4 minor 40 references

Bayesian data analysis for sky-averaged 21-cm experiments with contamination from linearly polarised foreground

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Polarised foreground leakage does not prevent recovery of the global 21-cm signal in most simulated cases.

desk verdict Useful simulation study, but the missing time-dependent polarization rotation means the pipeline is not actually validated for real observations. read the letter →

arxiv 2412.02552 v1 pith:OD5Z26TN submitted 2024-12-03 astro-ph.CO astro-ph.IM

Emma Shen , Dominic Anstey , Marta Spinelli , Eloy de Lera Acedo , Anastasia Fialkov This is my paper · ORCID
classification astro-ph.COastro-ph.IM
keywords global21-cmsignalCosmicDawnpolarisedforegroundsFaradayrotationBayesianinferenceStokesQleakagetime-separatedanalysisradiocosmology
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether leakage from linearly polarised Galactic foregrounds can destroy extraction of the global 21-cm absorption signal from Cosmic Dawn. The authors inject a Gaussian signal of about 0.16 K centred between 80 and 120 MHz into simulated antenna temperature data, contaminate the sky with three different models of Stokes $Q$ polarisation, and run REACH's time-separated Bayesian analysis pipeline. Their central claim is that the pipeline recovers the injected signal in the majority of cases, with root-mean-square error below 30% of the signal strength, whenever the polarisation fraction stays below roughly 3%. The reason to care is that polarised foregrounds rotated by Faraday effects create frequency-dependent structures that look like the signal; if the pipeline can separate them, one of the main known systematic threats to global 21-cm experiments is bounded.

What carries the argument

The machinery is a Bayesian pipeline that fits all time bins jointly. The likelihood in equation (17) gives each 5-minute time bin its own foreground model while sharing one set of global-signal parameters, and the fit is run with PolyChord nested sampling to compute evidences. Foregrounds are modelled physically by dividing the sky into regions of similar spectral index and convolving the scaled maps with the REACH dipole beam. The polarisation leakage itself is simulated through Stokes $Q$ maps whose Faraday rotation angle scales as $\lambda^2$, so its frequency structure is determined by the distribution of Faraday depths; that structure, and the antenna's chromatic response, are what the time-separated fit exploits to distinguish leakage from the signal.

What would settle it

Using a measured all-sky Stokes $Q$, $U$, and $V$ map at 50–150 MHz, inject a 0.157 K Gaussian signal into the simulated antenna temperature and run the time-separated pipeline; if the recovered-signal RMSE exceeds 30% of the injected amplitude at a polarisation fraction below $\sim 3\%$, the paper's stated bounds would not extend to the real sky.

Watch

Extended reading notes

Core claim

The discovery claim is that unaccounted-for linear polarisation contamination does not, in most simulated scenarios, prevent the time-separated REACH data analysis pipeline from reconstructing a 0.157 K Gaussian 21-cm absorption feature centred between 80 and 120 MHz. The paper shows this for polarised diffuse emission built by rotation measure synthesis, for an interpolated rotation-measure catalogue of point sources, and for toy maps with randomly assigned low Faraday depths. In the point-source case the signal is only missed at the highest polarisation fraction tested ($p=1/10$); it is recovered with low RMSE at $p=1/30$ and $1/70$. It also reports a structural result: contamination produced by the linear mixing of many Faraday-depth patches oscillates more rapidly and is easier to separate from the smooth signal than contamination from a single, slow oscillation, which becomes the hardest case.

Load-bearing premise

The load-bearing premise is that the three simulated Stokes $Q$ skies are representative of the real low-frequency polarized sky; real leakage could differ in spatial structure and Faraday-depth distribution, or include the Stokes $U$ and $V$ components these models omit.

Editorial extensions

If this is right

  • Below a polarisation fraction of roughly 3%, the pipeline recovers a 0.16 K signal without explicitly modelling polarisation leakage.
  • Polarised contamination that is spectrally fast, because it mixes many Faraday-depth components, is less dangerous than a single slow oscillatory component.
  • At $p=1/10$ the point-source model defeats the pipeline, placing an explicit bound on tolerable contamination.
  • These results give quantified performance bounds for a global 21-cm experiment against a class of known foreground systematics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the 3% threshold carries over to real data, experiments could monitor the sky's polarisation fraction rather than build leakage into the signal model; this is an extension, since the paper only tests simulated skies.
  • Because the simulations ignore Stokes $U$ and $V$ leakage, a testable next step is to inject full polarisation maps and see whether the recovered-signal RMSE remains below 30%.
  • The hardest real case implied here is a sky dominated by one slowly oscillating Faraday screen; targeting that structure with an extra foreground term may be the most effective safeguard.
  • A future all-sky polarisation survey at 50–150 MHz would let the same pipeline test be run on data-driven maps, converting the paper's bounds into a forecast for actual observations.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The manuscript studies whether the REACH time-separated Bayesian data analysis pipeline can recover an injected Gaussian global 21-cm signal (amplitude 0.157 K, width 10-15 MHz, centre 80-120 MHz) from simulated antenna temperature data contaminated by linearly polarized Galactic foregrounds. Three all-sky Stokes Q models are used: a rotation-measure-synthesis model based on the MWA survey, an interpolated rotation-measure catalogue (RMTable2023), and toy models with random low Faraday-depth regions. The pipeline jointly fits all time bins with physically motivated foreground models using PolyChord, and recovery is evaluated with a signal RMSE criterion (good fit if RMSE < 30% of the injected amplitude). The paper reports successful recovery in most tested cases for polarization fractions below roughly 3%, and concludes that the REACH pipeline is robust to this class of contamination.

Significance. If correct, the result would be a valuable validation of the REACH pipeline against a chromatic systematic that is known to degrade standard smooth-foreground analyses. The paper's main strengths are its explicit injection-recovery design, the use of three qualitatively distinct polarized foreground models, the physically informed foreground and beam modelling, and the clear RMSE-based success metric; the injected signal is not used as a prior, so the test is not circular. The principal limitation is that the simulated leakage is not a faithful model of what a real fixed dipole measures, which makes the quantitative recovery claim conditional on a forward-model upgrade rather than directly applicable to the physical observation.

major comments (3)
  1. [Section 2.1, Eqs. (5)-(6); Appendix A, Eqs. (A3)-(A7)] The forward model omits the rotation of Stokes Q into the antenna polarization frame and omits Stokes U (and V). A linearly polarized dipole responds to E_xx proportional to I + Q', with Q' = Q cos(2 chi) + U sin(2 chi), where chi is the position angle of the antenna polarization axis on the sky and varies with time as the sky rotates. In the simulation, a sky-frame Stokes Q map is added directly to T_f and then convolved with a scalar power pattern, so the leakage has no time-dependent parallactic-angle modulation. Because the time-separated pipeline is specifically designed to exploit time-varying chromatic structure, this omission is load-bearing for the central recovery claim. I recommend replacing the scalar-beam treatment with a Jones-matrix forward model (at minimum a time-dependent Q' rotation; ideally including U and V), or explicitly restricting the conclusions to the idealized co-aligned static case.
  2. [Section 4.2, Figs. 4-5, Section 5, Abstract] The polarization-fraction thresholds are internally inconsistent. The text states that 'signal RMSE of less than 10% ... can be achieved for linear polarisation fractions of 1/30 and 1/10', while Fig. 4 shows the p = 1/10 case completely missing the signal; Section 5 states a threshold of '<=3%' even though p = 1/30 is 3.3%; and the Abstract says 'below ~3%'. The exact p values that satisfy the 30% and 10% RMSE criteria should be reconciled and stated consistently across the abstract, results, and conclusions.
  3. [Section 4, Eq. (18), Figs. 5 and 8] The reported RMSE values are based on a single noise realization per configuration, and the adopted noise level sigma_n is not stated. Because the paper's quantitative success criterion is RMSE < 30% of the injected signal strength, the absence of any noise-realization scatter (or at least the value of sigma_n) leaves the 'all tested cases' claim without a statistical uncertainty. Repeating the analysis over multiple noise draws, or stating sigma_n and showing representative stability, would strengthen the central claim.
minor comments (4)
  1. [Section 3.1 and throughout] There are several typographical issues, including the duplicated phrase 'parameters parameters' in Section 3.1, 'theflattened Gaussian' in the Introduction, and a stray quotation mark after 'nearest-neighbour interpolation.' in Section 5.
  2. [Section 5] The statement that the REACH pipeline 'significantly outperforms standard approaches' is not directly demonstrated by this paper, since no standard pipeline is run on the same simulated data; the support comes only from cited earlier work. Either add a baseline comparison or soften the claim.
  3. [Fig. 5 and Section 4.2] The caption of Fig. 5 and the text of Section 4.2 disagree on the injected-signal width used for the RMSE curves (10 MHz in the caption versus 15 MHz in the text); please make the figure-specific parameters explicit.
  4. [Section 2.1, Eq. (6)] The time dependence of the simulated sky/beam is not written explicitly in Eq. (6), although the pipeline is time-separated; a sentence clarifying how the 5-minute time bins are produced would help reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the pipeline recovery claim is benchmarked against externally injected simulations, not against its own inputs.

full rationale

The paper's central claim is an injection-recovery test: synthetic antenna temperatures are generated from Eqs. (5)-(6) with a known injected Gaussian 21-cm signal T_21 and an unmodeled Stokes-Q leakage T_Q, and the REACH pipeline's reconstruction T_S is compared to T_21 via the RMSE of Eq. (18). The injected signal enters the analysis only as the target of evaluation; its parameters are not used as priors or constraints. The pipeline's free parameters (foreground spectral indices, signal parameters, noise) are estimated jointly by PolyChord from the simulated data, and the signal prior is described as 'a standard global 21-cm signal within 3 sigmas' (Fig. 6 caption), not as the injected values. The recovery is nontrivial: at p=1/10 in the point-source model (Section 4.2) the pipeline 'completely misses the signal,' which confirms the test is not forced. Self-citations to Anstey et al. (2021, 2023) for the REACH pipeline are references to an existing analysis code used as the object of study, and the conclusions are supported by the simulations here rather than by those citations alone. The acknowledged simplifications - Stokes Q only, no U/V, Faraday-depth cutoff, no parallactic-angle rotation (Eq. 5, Section 2.2, Appendix A) - are external-validity limitations for real polarized skies and would motivate a corrected-simulation concern about the physical realism of the test scenario, but they do not make the pipeline's recovery logically depend on its own inputs. No step of the claimed derivation chain is equivalent to its inputs by construction.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central claim rests on a small set of modeling assumptions rather than derived physical laws. The free parameters listed are simulation inputs chosen to explore a parameter space, not fitted constants. No new entities are introduced. The most consequential assumptions are the representative nature of the polarized sky models and the neglect of Stokes U and V leakage, both acknowledged by the authors.

free parameters (3)
  • Uniform linear polarisation fraction p = 1/10, 1/30, 1/70, 1/1000, 1/3000 (simulation inputs)
    Chosen by hand to probe the pipeline's tolerance; the claimed ~3% threshold is inferred from these discrete values, not fitted to data.
  • Faraday depth cutoff psi_max = 5 rad/m^2
    Excluded higher Faraday depths based on cited observations (Haverkorn et al. 2004; Bernardi et al. 2009; Lenc et al. 2016); this cutoff affects the oscillation rate of contamination in the rotation measure synthesis model.
  • Random Faraday depth range in toy model = 0 to 0.5 rad/m^2
    Chosen to produce slow oscillations resembling the injected signal, making the test deliberately challenging. The range is an ad hoc modeling choice.
assumptions (6)
  • standard math Faraday rotation of linearly polarized synchrotron emission follows chi(lambda^2) = chi0 + psi lambda^2, with Faraday depth psi.
    Used in Section 2.2 and equation (7) to generate the frequency dependence of Stokes Q contamination; this is the standard physical relation for Faraday rotation.
  • domain assumption Stokes U and V leakage are negligible for the REACH dipole receiver; only Stokes Q contributes to the leakage term T_Q.
    Appendix A derives T_xx = T_f + T_Q + T_21 for a single linearly polarized receptor (equations A4-A7). Real receivers can have cross-polarization and V leakage, which are not modelled.
  • domain assumption The polarized diffuse emission model based on the MWA 2400 deg^2 survey at 189 MHz, extrapolated to all-sky and to 50-150 MHz, is representative of the true low-frequency polarized sky.
    Section 2.2.1 states observations at the relevant frequencies are lacking and the simulation is an extrapolation; if the real sky has different Faraday depth structure, the contamination spectra will differ.
  • domain assumption The RMTable2023 point-source catalogue with nearest-neighbour interpolation provides a plausible all-sky Faraday depth map.
    Section 2.2.2: the catalogue does not cover the whole sky and is of discrete sources; interpolating it to a continuous map is an ad hoc choice that affects the simulated Stokes Q structure.
  • ad hoc to paper The global 21-cm signal is well approximated by a Gaussian of amplitude 0.157 K and width 10-15 MHz for the purpose of this test.
    Stated in Section 2.1 and used for all injections; the real signal shape is uncertain and may deviate from a Gaussian, affecting how easily it can be separated from smooth foregrounds.
  • domain assumption The REACH foreground model, which divides the sky into regions of uniform spectral index and fits the region spectral indices, is flexible enough to absorb the simulated foreground but not the injected signal.
    This is the core modelling assumption of the pipeline (Section 3.2, Anstey et al. 2021, 2023); the paper's recovery results depend on this balance holding for the tested cases.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Bayesian data analysis for sky-averaged 21-cm experiments with contamination from linearly polarised foreground." pith.science (2026). https://pith.science/paper/OD5Z26TN

@misc{pith2026241202552,
  author       = {Pith},
  title        = {Pith review of: Bayesian data analysis for sky-averaged 21-cm experiments with contamination from linearly polarised foreground},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OD5Z26TN}},
  note         = {Machine review of arXiv:2412.02552}
}
abstract

The precise measurement of the sky-averaged HI absorption signal between 50 and 200 MHz is the primary goal of global 21-cm cosmology. This measurement has the potential to unravel the underlying physics of cosmic structure formation and evolution during the Cosmic Dawn. It is, however, hindered by various non-smooth, frequency-dependent effects, whose structures resemble those of the signal. One such effect is the leakage of polarised foregrounds into the measured intensity signal: polarised foreground emission undergoes Faraday rotation as it passes through the magnetic fields of the interstellar medium, imprinting a chromatic structure in the relevant frequency range which complicates the extraction of the cosmological HI absorption feature. We investigate the effect of polarised Galactic foregrounds on extracting the global 21-cm signal from simulated data using REACH's data analysis pipeline; the Radio Experiment for the Analysis of Cosmic Hydrogen (REACH) is an experiment designed to detect the sky-averaged 21-cm HI signal from the early Universe using physically informed models. Using the REACH pipeline, we successfully recover an injected global 21-cm signal with an amplitude of approximately 0.16 K, centred between 80 and 120 MHz, achieving a low root-mean-square error (less than 30\% of the injected signal strength) in all the tested cases. This includes scenarios with simulated polarised Galactic diffuse emissions and polarised point source emissions, provided the overall polarisation fraction is below $\sim 3\%$. The linear mixing of contamination, caused by the superposition of multiple patches with varying strengths of Faraday rotation, produces patterns that are more distinct from the global signal. This distinction makes global signal recovery easier compared to contamination resulting from a single, slow oscillation pattern.

Figures

Figures reproduced from arXiv: 2412.02552 by the authors.

Figure 1
Figure 1. The all-sky distribution of Stokes Q or Faraday depth value of the three different adopted models. The maps are in the Galactic coordinates. Top: the map on the top of the figure is the Stokes 𝑄 map generated by the polarised Galactic diffuse emission simulation described in Spinelli et al. (2018) and section 2.2.1. In this model, higher values of Faraday depth, 𝜓 > 5 rad m−2 , are excluded as observations have show… view at source ↗
Figure 2
Figure 2. The reconstructed signal (blue contours) and foreground residual (red contours) plots of an example case by the rotation measure synthesis technique. The yellow dashed line represents the injected signal with an amplitude of 0.157 K centred at 90 MHz. This example considers only low Faraday depth values (𝜓 ≤ 5 rad m−2 ). Visually, the reconstructed signal contours are already very close to the injected signal. The r… view at source ↗
Figure 3
Figure 3. Scaled Stokes 𝑄 as a function of frequency. A uniform linear polarisation fraction of 𝑝 = 1/10 is applied to the whole sky at all frequencies. The Stokes 𝑄 parameter map is simulated according to equation (12), and the Faraday depth is based on the RMTable2023 catalogue (Van Eck et al. 2023). The Stokes 𝑄 structure in blue solid line is rather noisy compared to the injected global signals of 15 MHz width in yellow d… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Reconstructed signal and foreground residual plots for cases with spatially uniform linear polarisation fraction 𝑝 of 1/10, 1/30, and 1/70 from left to right respectively. The same polarisation fraction is applied to the whole sky at all frequencies in each case. The F…
Figure 5
Figure 5. Figure 5: The signal RMSE for cases with different polarisation fractions and injected signals of amplitude 0.157 K and width 10 MHz, centred at 80, 100, and 120 MHz respectively. As expected, signal recovery significantly improves with decreasing polarisation fraction across al…
Figure 7
Figure 7. Figure 7: Reconstructed signal and foreground residual plots for cases with 2, 4, and 25 Faraday regions in which Faraday depth values are randomly assigned between 0 and 0.5 rad m−2 . The regions are divided based on the spectral indices across the sky. The homogeneous linear p…
Figure 8
Figure 8. Figure 8: The signal RMSE for cases with different numbers of Faraday depth regions for an injected signal of amplitude 0.157 K, width 15 MHz, and centred at 90 MHz. The regions are divided based on the spectral index across the sky, as described in Anstey et al. (2021). The sig…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 4 canonical work pages

  1. [1]

    Anstey D., de Lera Acedo E., Handley W., 2021, @doi [ ] 10.1093/mnras/stab1765 , https://ui.adsabs.harvard.edu/abs/2021MNRAS.506.2041A 506, 2041

    doi:10.1093/mnras/stab1765 2021

  2. [2]

    Anstey D., de Lera Acedo E., Handley W., 2023, @doi [ ] 10.1093/mnras/stad156 , https://ui.adsabs.harvard.edu/abs/2023MNRAS.520..850A 520, 850

    doi:10.1093/mnras/stad156 2023

  3. [3]

    Barkana R., Loeb A., 2001, @doi [ ] 10.1016/S0370-1573(01)00019-9 , https://ui.adsabs.harvard.edu/abs/2001PhR...349..125B 349, 125

    doi:10.1016/s0370-1573(01)00019-9 2001

  4. [4]

    Bernardi G., et al., 2009, @doi [ ] 10.1051/0004-6361/200911627 , https://ui.adsabs.harvard.edu/abs/2009A&A...500..965B 500, 965

    doi:10.1051/0004-6361/200911627 2009

  5. [5]

    Bernardi G., et al., 2013, @doi [ ] 10.1088/0004-637X/771/2/105 , https://ui.adsabs.harvard.edu/abs/2013ApJ...771..105B 771, 105

    doi:10.1088/0004-637x/771/2/105 2013

  6. [6]

    Bevins H. T. J., Handley W. J., Fialkov A., de Lera Acedo E., Greenhill L. J., Price D. C., 2020, arXiv e-prints, https://ui.adsabs.harvard.edu/abs/2020arXiv200714970B p. arXiv:2007.14970

    work page Pith review arXiv 2020

  7. [7]

    D., Rogers A

    Bowman J. D., Rogers A. E. E., Hewitt J. N., 2008, @doi [ ] 10.1086/528675 , https://ui.adsabs.harvard.edu/abs/2008ApJ...676....1B 676, 1

    doi:10.1086/528675 2008

  8. [8]

    D., Rogers A

    Bowman J. D., Rogers A. E. E., Monsalve R. A., Mozdzen T. J., Mahesh N., 2018, @doi [Nature] 10.1038/nature25792 , https://ui.adsabs.harvard.edu/abs/2018Natur.555...67B 555, 67

    doi:10.1038/nature25792 2018

Show all 40 references
  1. [9]

    J., 1966, @doi [ ] 10.1093/mnras/133.1.67 , https://ui.adsabs.harvard.edu/abs/1966MNRAS.133...67B 133, 67

    Burn B. J., 1966, @doi [ ] 10.1093/mnras/133.1.67 , https://ui.adsabs.harvard.edu/abs/1966MNRAS.133...67B 133, 67

    1966 doi

  2. [10]

    arXiv:2103.05085

    Burns J., et al., 2021, @doi [arXiv e-prints] 10.48550/arXiv.2103.05085 , https://ui.adsabs.harvard.edu/abs/2021arXiv210305085B p. arXiv:2103.05085

    2021 doi arXiv

  3. [11]

    Carretti E., et al., 2019, @doi [ ] 10.1093/mnras/stz806 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.489.2330C 489, 2330

    2019 doi

  4. [12]

    J., et al

    Chrysostomou A., Ball L., Diamond P. J., et al. 2020, SKA1 Operations Plan

    2020

  5. [13]

    A., Hutschenreuter S., Vacca V., Oppermann N., 2017, @doi [Phys

    En lin T. A., Hutschenreuter S., Vacca V., Oppermann N., 2017, @doi [Phys. Rev. D] 10.1103/PhysRevD.96.043021 , 96, 043021

    2017 doi

  6. [14]

    B., 1958, @doi [Proceedings of the IRE] 10.1109/JRPROC.1958.286741 , https://ui.adsabs.harvard.edu/abs/1958PIRE...46..240F 46, 240

    Field G. B., 1958, @doi [Proceedings of the IRE] 10.1109/JRPROC.1958.286741 , https://ui.adsabs.harvard.edu/abs/1958PIRE...46..240F 46, 240

    1958

  7. [15]

    R., Oh S

    Furlanetto S. R., Oh S. P., Briggs F. H., 2006, @doi [ ] 10.1016/j.physrep.2006.08.002 , https://ui.adsabs.harvard.edu/abs/2006PhR...433..181F 433, 181

    2006 doi

  8. [16]

    J., Hobson M

    Handley W. J., Hobson M. P., Lasenby A. N., 2015, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stv1911 , 453, 4384

    2015 doi

  9. [17]

    G., 2004, @doi [ ] 10.1051/0004-6361:200400051 , https://ui.adsabs.harvard.edu/abs/2004A&A...427..549H 427, 549

    Haverkorn M., Katgert P., de Bruyn A. G., 2004, @doi [ ] 10.1051/0004-6361:200400051 , https://ui.adsabs.harvard.edu/abs/2004A&A...427..549H 427, 549

    2004 doi

  10. [18]

    D., Puchwein E., 2018, @doi [Nature] 10.1038/s41586-018-0796-5 , https://ui.adsabs.harvard.edu/abs/2018Natur.564E..32H 564, E32

    Hills R., Kulkarni G., Meerburg P. D., Puchwein E., 2018, @doi [Nature] 10.1038/s41586-018-0796-5 , https://ui.adsabs.harvard.edu/abs/2018Natur.564E..32H 564, E32

    2018 doi

  11. [19]

    J., Rees M

    Hogan C. J., Rees M. J., 1979, @doi [ ] 10.1093/mnras/188.4.791 , https://ui.adsabs.harvard.edu/abs/1979MNRAS.188..791H 188, 791

    1979 doi

  12. [20]

    Lenc E., et al., 2016, @doi [ ] 10.3847/0004-637X/830/1/38 , https://ui.adsabs.harvard.edu/abs/2016ApJ...830...38L 830, 38

    2016 doi

  13. [21]

    J., 1997, @doi [ ] 10.1086/303549 , https://ui.adsabs.harvard.edu/abs/1997ApJ...475..429M 475, 429

    Madau P., Meiksin A., Rees M. J., 1997, @doi [ ] 10.1086/303549 , https://ui.adsabs.harvard.edu/abs/1997ApJ...475..429M 475, 429

    1997 doi

  14. [22]

    A., et al., 2024, @doi [ ] 10.1093/mnras/stae1138 , https://ui.adsabs.harvard.edu/abs/2024MNRAS.530.4125M 530, 4125

    Monsalve R. A., et al., 2024, @doi [ ] 10.1093/mnras/stae1138 , https://ui.adsabs.harvard.edu/abs/2024MNRAS.530.4125M 530, 4125

    2024 doi

  15. [23]

    Patra N., Subrahmanyan R., Raghunathan A., Udaya Shankar N., 2013, @doi [Experimental Astronomy] 10.1007/s10686-013-9336-3 , https://ui.adsabs.harvard.edu/abs/2013ExA....36..319P 36, 319

    2013 doi

  16. [24]

    Philip L., et al., 2019, @doi [Journal of Astronomical Instrumentation] 10.1142/S2251171719500041 , https://ui.adsabs.harvard.edu/abs/2019JAI.....850004P 8, 1950004

    2019 doi

  17. [25]

    C., et al., 2018, @doi [ ] 10.1093/mnras/sty1244 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.478.4193P 478, 4193

    Price D. C., et al., 2018, @doi [ ] 10.1093/mnras/sty1244 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.478.4193P 478, 4193

    2018 doi

  18. [26]

    R., Loeb A., 2008, @doi [ ] 10.1103/PhysRevD.78.103511 , https://ui.adsabs.harvard.edu/abs/2008PhRvD..78j3511P 78, 103511

    Pritchard J. R., Loeb A., 2008, @doi [ ] 10.1103/PhysRevD.78.103511 , https://ui.adsabs.harvard.edu/abs/2008PhRvD..78j3511P 78, 103511

    2008 doi

  19. [27]

    B., Lightman A

    Rybicki G. B., Lightman A. P., 1986, Radiative Processes in Astrophysics . Wiley-VCH

    1986

  20. [28]

    H., Pober J

    Sims P. H., Pober J. C., 2020, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stz3388 , https://ui.adsabs.harvard.edu/abs/2020MNRAS.492...22S 492, 22

    2020 doi

  21. [29]

    Singh S., Subrahmanyan R., 2019, @doi [ ] 10.3847/1538-4357/ab2879 , https://ui.adsabs.harvard.edu/abs/2019ApJ...880...26S 880, 26

    2019 doi

  22. [30]

    Singh S., et al., 2017, @doi [ ] 10.3847/2041-8213/aa831b , https://ui.adsabs.harvard.edu/abs/2017ApJ...845L..12S 845, L12

    2017 doi

  23. [31]

    Singh S., et al., 2022, @doi [Nature Astronomy] 10.1038/s41550-022-01610-5 , https://ui.adsabs.harvard.edu/abs/2022NatAs...6..607S 6, 607

    2022 doi

  24. [32]

    S., Skilling J., 2006, Data Analysis - A Bayesian Tutorial , 2nd edn

    Sivia D. S., Skilling J., 2006, Data Analysis - A Bayesian Tutorial , 2nd edn. Oxford Science Publications, Oxford University Press

    2006

  25. [33]

    Skilling J., 2006, @doi [Bayesian Analysis] 10.1214/06-BA127 , 1, 833

    2006 doi

  26. [34]

    G., 2018, @doi [ ] 10.1093/mnras/sty1457 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.479..275S 479, 275

    Spinelli M., Bernardi G., Santos M. G., 2018, @doi [ ] 10.1093/mnras/sty1457 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.479..275S 479, 275

    2018 doi

  27. [35]

    G., 2019, @doi [ ] 10.1093/mnras/stz2425 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.489.4007S 489, 4007

    Spinelli M., Bernardi G., Santos M. G., 2019, @doi [ ] 10.1093/mnras/stz2425 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.489.4007S 489, 4007

    2019 doi

  28. [36]

    L., et al., 2019, @doi [ ] 10.1051/0004-6361/201834777 , https://ui.adsabs.harvard.edu/abs/2019A&A...623A..71V 623, A71

    Van Eck C. L., et al., 2019, @doi [ ] 10.1051/0004-6361/201834777 , https://ui.adsabs.harvard.edu/abs/2019A&A...623A..71V 623, A71

    2019 doi

  29. [37]

    L., et al., 2023, @doi [ ] 10.3847/1538-4365/acda24 , https://ui.adsabs.harvard.edu/abs/2023ApJS..267...28V 267, 28

    Van Eck C. L., et al., 2023, @doi [ ] 10.3847/1538-4365/acda24 , https://ui.adsabs.harvard.edu/abs/2023ApJS..267...28V 267, 28

    2023 doi

  30. [38]

    A., 1952, @doi [ ] 10.1086/106661 , https://ui.adsabs.harvard.edu/abs/1952AJ.....57R..31W 57, 31

    Wouthuysen S. A., 1952, @doi [ ] 10.1086/106661 , https://ui.adsabs.harvard.edu/abs/1952AJ.....57R..31W 57, 31

    1952 doi

  31. [39]

    de Lera Acedo E., et al., 2022, @doi [Nature Astronomy] 10.1038/s41550-022-01709-9 , https://ui.adsabs.harvard.edu/abs/2022NatAs...6..984D 6, 984

    2022 doi

  32. [41]

    M., Jonas J., Landecker T

    de Oliveira-Costa A., Tegmark M., Gaensler B. M., Jonas J., Landecker T. L., Reich P., 2008b, @doi [ ] 10.1111/j.1365-2966.2008.13376.x , https://ui.adsabs.harvard.edu/abs/2008MNRAS.388..247D 388, 247

    2008

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.