Pith. sign in

REVIEW 3 major objections 4 minor 47 references

No 21-cm absorption is detected toward the z = 5.84 quasar J2329-1520, placing a 95% upper limit of τ21 < 0.02 on the intergalactic medium over 5.38 < z < 5.84.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 10:02 UTC pith:KWBKGHUP

load-bearing objection A careful null detection that yields a genuinely new upper limit on 21-cm optical depth at z~5.6, but the abstract overstates what it proves about IGM heating. the 3 major comments →

arxiv 2607.20371 v1 pith:KWBKGHUP submitted 2026-07-22 astro-ph.CO astro-ph.GA

The burgeoning rise of the 21-cm forest I: constraints on the optical depth of the intergalactic medium at 5.38 < z < 5.84

Chanasorn Kongprachaya , Gianni Bernardi , Emilio Ceccotti , Andrei Mesinger , Benedetta Ciardi , Oleg M. Smirnov This is my paper
classification astro-ph.CO astro-ph.GA
keywords 21-cm forestEpoch of Reionization21-cm absorptionintergalactic mediumspin temperatureoptical depthradio-loud quasarlow-frequency radio observations
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper tries to measure how much neutral hydrogen remains along the line of sight to a bright radio quasar at the end of reionization, using the 21-cm line in absorption against the quasar's radio continuum. Analyzing 18 hours of low-frequency radio observations, the authors find no sign of such absorption. If their interpretation is right, the intergalactic medium along this sightline is not filled with cold, neutral patches: the 21-cm optical depth is below about 0.02 and any residual neutral island must be warmer than 1.73 K. That matters because it gives an independent, direct probe of IGM heating near the end of reionization, complementing 21-cm power-spectrum limits and disfavoring very cold hydrogen at z < 5.84.

Core claim

The paper establishes upper limits on the 21-cm optical depth of the intergalactic medium at 5.38 < z < 5.84 by fitting a power-law quasar continuum modified by 21-cm absorption. Under the global model, where neutral fraction and spin temperature evolve smoothly with redshift, the 95% confidence upper limit is τ21 < 0.02, ranging from 2.6 × 10⁻² at z = 5.84 to 1.5 × 10⁻² at z = 5.38. Under the island model, where a single cold neutral patch survives, the authors set a 95% lower limit on the spin temperature of any residual HI region of T_s > 1.73 K, assuming mean density. These limits exceed the adiabatic cooling floor of roughly 0.74–0.85 K, providing independent evidence that the IGM was h

What carries the argument

The central machinery is the standard 21-cm optical depth formula, τ21 ≈ 0.0092 (1 + δ)/T_s · x_HI (1 + z)^{3/2}, combined with a power-law continuum model S(ν) = S_200 (ν/ν_*)ᵅ e^{−τ21(ν)}. Two absorption scenarios are tested: a global model with a tanh redshift evolution of neutral fraction and a log-interpolated spin temperature, and an island model with a rectangular optical-depth patch representing a cold neutral island. The continuum and absorption parameters are fitted jointly via MCMC, using a covariance matrix estimated from residual image cubes rather than assuming independent channels.

Load-bearing premise

The search assumes that J2329-1520's intrinsic radio spectrum is a single, temporally constant power law across the observed band; if the spectrum curves within 203–222.5 MHz or the quasar varied intrinsically between the 2018 and 2022 runs, the fitted continuum could hide or mimic an absorption signal.

What would settle it

Compare spectra from the 2018 and 2022 epochs separately after calibration: if the residual spectral shapes differ beyond the estimated noise, the constant power-law assumption fails. Alternatively, a deeper observation with sub-mJy channel noise that resolves narrow 21-cm forest lines — or a detection of 21-cm absorption toward another z ≈ 5.5–6 quasar — would directly test the null detection.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • The 21-cm optical depth along this quasar sightline is below about 0.02 at 95% confidence over 5.38 < z < 5.84, tightening constraints on neutral hydrogen near the end of reionization.
  • Residual neutral islands, if they exist, must have spin temperature above 1.73 K, so extremely cold HI regions are disfavored at z < 5.84.
  • The IGM was heated above the adiabatic cooling limit of roughly 0.8 K at these redshifts, consistent with theoretical expectations and with higher-redshift 21-cm power-spectrum measurements.
  • The joint continuum-plus-absorption fitting approach, validated on simulated spectra, can be applied to other high-redshift radio quasars to build a statistical sample of 21-cm absorption constraints.
  • The measured quasar continuum, S_200 = 86.8 ± 1.1 mJy with spectral index −0.88 ± 0.02 from 150 MHz to 3 GHz, provides a stable reference for future absorption searches toward this source.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • A single sightline cannot distinguish a globally warm IGM from one with patchy cold gas along this particular direction; observing several quasars at comparable redshifts would test whether the T_s > 1.73 K constraint is universal or line-of-sight dependent.
  • If the upper limit tightens with future deep low-frequency observations, stacking the growing sample of z > 5.5 radio-loud quasars could push τ21 constraints toward the ~10⁻⁴ level where 21-cm forest structure becomes detectable.
  • The same modeling framework could be extended to lower observed frequencies to probe redshifts beyond z = 5.84, reaching the more neutral pre-reionization IGM where absorption is expected to be stronger.
  • The discrepancy between the measured per-channel noise (3.6 mJy/beam) and the radiometer expectation (about 2.4 mJy/beam from Stokes V) hints that residual systematic noise remains; reducing it would directly improve the τ21 limit.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper presents 18 hours of uGMRT observations of the radio-loud quasar PSO J352.4034-15.3373 (J2329-1520, z=5.84) in the 203-222.5 MHz band, together with archival measurements from 150 MHz to 3 GHz. After detailed calibration, self-calibration, and joint imaging, the authors extract a 50-channel spectrum with 3.6 mJy/beam rms per 390 kHz channel. They jointly fit a power-law continuum plus two 21-cm absorption models: a 'global' model with a tanh reionization history and log-linear spin-temperature evolution, and an 'island' model with a rectangular absorption profile from a fully neutral patch. They report no absorption: under the island model they derive Ts > 1.73 K (95% C.L., assuming delta=0), and under the global model they derive tau21 < 0.02 (95% C.L., more precisely in the range 1.5e-2 to 2.6e-2 over z=5.38-5.84). Simulations with the same pipeline are used to validate parameter recovery for brighter sources and for J2329-1520. The paper concludes that the IGM was heated above the adiabatic cooling limit at z ~ 5.7 and that extremely cold HI islands are disfavoured.

Significance. If the null detection and limits hold, this is one of the first direct 21-cm absorption constraints on the IGM near the end of reionization from a high-redshift radio quasar, complementing 21-cm power-spectrum upper limits. The paper has notable strengths: the calibration and flagging workflow is detailed; the likelihood uses a full covariance matrix estimated from the residual image cube; an independent Stokes V noise estimate is provided; and the continuum and absorption parameters are fitted jointly rather than in two separate steps. The simulations and MCMC setup are clearly described. However, the headline conclusions are more model-dependent than the abstract suggests, and the central tau21 upper limit relies on an untested assumption that the source continuum is an unbroken power law stable across epochs and within the band. These issues require additional analysis before the constraints can be considered fully load-bearing.

major comments (3)
  1. [§3, Eq. (1)] The central upper limit rests on the assumption that the intrinsic quasar spectrum is a single power law that is constant across all epochs. A 95% upper limit of tau21 ~ 0.02 corresponds to only ~1.7 mJy against an ~87 mJy continuum, i.e., roughly half the per-channel rms; the tight limit is obtained by jointly fitting about 50 channels. The residual-image covariance matrix (Eq. 7) captures thermal noise and pixel correlations, but not continuum model error. In addition, the amplitude self-calibration step in §3 uses 25 MHz frequency solution intervals and can in principle absorb smooth spectral structure. No test is shown for power-law curvature, inter-epoch source variability, or consistency between the 2018 and 2022 runs. I request explicit robustness tests: (i) refit with an extra curvature term (e.g., a log-polynomial) and report how the tau21 limit changes; (ii) fit the 2018 and 20
  2. [Abstract and §4.2] The Abstract's statement that the results 'confirm that the IGM was heated above the adiabatic cooling limit' is stronger than the data support. The Ts > 1.73 K lower limit is derived only under the island model with xHI = 1 and delta = 0. If the absorbing patch is partially ionized or underdense, the implied lower limit on Ts weakens, because tau21 is proportional to xHI(1+delta)/Ts. It can then fall below the adiabatic range 0.74-0.85 K quoted in Eq. (8). Moreover, this limit applies to residual neutral islands, not to the bulk IGM. I recommend replacing 'confirm' with a more cautious formulation, such as 'provide a lower limit consistent with heating above the adiabatic floor under the island-model assumptions,' and stating the xHI = 1, delta = 0 assumptions in the Abstract.
  3. [§4.1] The paragraph stating that the observed constraints are 'consistent with what is presented in the simulated case, indicating a negligible contribution from systematic errors' is not a valid systematic-error test. The simulations use the same power-law-plus-absorption model and the same Gaussian likelihood as the real-data fit; agreement only demonstrates that the MCMC recovers the input model under idealised noise. It does not probe calibration errors, continuum model misspecification, RFI residuals, or epoch-dependent gain errors. This sentence should be removed or replaced with the empirical consistency checks described in the first major comment.
minor comments (4)
  1. [§4, first paragraph] The text refers to 'the J2329-1530 source', but the source is named J2329-1520 elsewhere. Please correct the typo.
  2. [Abstract and §4] The text states a final spectrum spanning 203-222.5 MHz with 50 channels, while the quoted absorption redshift range is 5.38 < z < 5.84, which corresponds to 207.6-222.5 MHz. Please clarify exactly which channels enter the fits and whether data below the source redshift line (203-207.6 MHz) are used. If they are used, justify why (e.g., source-associated absorption or redshift uncertainty) and show that the results are unchanged when they are excluded.
  3. [Table 2] The entry for CFHQS J142952+544717 is abbreviated as 'J1429+5547', which is inconsistent with the source name (should be J1429+5447 or similar). Please check the shorthand notation.
  4. [Appendix A, Table A.4] The notes for J1427+3312 list two references as '(a)'; one should be re-lettered to avoid ambiguity.

Circularity Check

0 steps flagged

No significant circularity: the τ21<0.02 and Ts>1.73 K limits are null-detection bounds from the joint fit of Eq. (1) to the observed spectrum; no fitted parameter is renamed as a prediction, and self-citations are methodological, not load-bearing.

full rationale

Walking the derivation chain: the observed spectrum (Fig. 17) is fit with Eq. (1), a power-law continuum multiplied by exp(-τ21), with continuum parameters (S200, α) and 21-cm parameters (T⋆, zr, Δz for the global model; (1+δ)/Ts, νp, Δν for the island model) sampled jointly. The headline claims are not predictions from fitted constants: they are 95% upper/lower bounds obtained because the absorption parameters are consistent with zero. The paper states the best fit is 'a power law continuum with no absorption' (§4.2), the τ21<0.02 limit is the derived 95% C.I. of the fitted optical depth profile (Fig. 20), and Ts>1.73 K is the transform of the (1+δ)/Ts posterior (Fig. 21), which lies well inside its uniform prior [0,2] K^-1, so the limit is data-driven rather than prior-forced. The T_s(z) and x_HI(z) shapes are modeling assumptions that the constraint inherits, but the target quantities are free parameters, so the model is not defined in terms of the result. Self-citations (Bernardi et al. 2013 for the covariance estimation of Eq. 7; Ceccotti et al. 2025; Spinelli et al. 2019) are methodological or contextual, and the covariance method is also anchored externally (de Oliveira-Costa et al. 2008); none is invoked to forbid alternatives or supply the result. The injection-recovery simulations (§2.3) are labeled as method validation, and the 'consistent with the simulated case' remark in §4.1 is a self-consistency check, not an external benchmark. The paper's own caveats (the 'simplified, toy model' for temperature evolution in §2.1; the 50% noise discrepancy between the Stokes V and Stokes I estimates in §4) are honest limitations about systematics, not circular steps. The power-law continuum and constant-source assumptions are correctness risks (unmodeled spectral curvature or inter-epoch variability could bias the quoted limit), but that is model-error sensitivity, not a reduction of the result to its own inputs.

Axiom & Free-Parameter Ledger

8 free parameters · 8 axioms · 0 invented entities

The central result is a model-dependent upper limit. It relies on standard radio cosmology plus the paper's own toy models for x_HI and T_s. No new physical entities are introduced.

free parameters (8)
  • S_200 (flux density at 200 MHz) = 86.8 ± 1.1 mJy (1σ; 95% C.I. [84.73, 89.00] mJy)
    Continuum normalization fitted jointly to uGMRT spectrum and archival flux measurements (Eq. 1); the absorption search is a fractional deviation from this continuum, so the fitted S_200 is load-bearing.
  • α (spectral index) = -0.88 ± 0.02
    Power-law slope fitted to the spectrum; any curvature or error in α changes residual wiggles that could mimic or hide absorption.
  • T⋆ (spin temperature at z=6) = Weakly constrained; posterior not quoted separately
    Global model parameter controlling the amplitude of τ21(z) via Eq. (4); sampled through the combination (1+δ)/T_s.
  • z_r (reionization midpoint) = Posterior, uniform prior [0,9]
    Tanh model for x_HI(z); determines where the optical depth peaks.
  • Δz (reionization duration) = Posterior, uniform prior [0,9]
    Width of the x_HI transition; broadens/narrows the τ21(z) profile.
  • (1+δ)/T_s island ratio = Upper limit corresponds to T_s>1.73 K at 95% (δ=0)
    Island model amplitude; fitted to the spectrum with uniform prior [0,2] K^-1.
  • νp (island central frequency) = Uniform over [207.6,222.5] MHz
    Position of the hypothetical absorption feature; unconstrained by data.
  • Δν (island width) = Uniform over [0.39,5] MHz
    Width of the rectangular absorption; unconstrained apart from priors.
axioms (8)
  • standard math Standard 21-cm optical depth formula, Eq. (2), with prefactor 0.0092 and the assumption dv∥/dr∥=H/(1+z).
    Used for both models; standard result from Furlanetto et al. (2006).
  • domain assumption The intrinsic quasar spectrum is a single power law from 150 MHz to 3 GHz.
    Eq. (1); the no-absorption conclusion is relative to a power-law continuum.
  • domain assumption The line of sight is at mean density δ=0.
    Used in the T_s lower-limit conversion and in the τ21 formula; stated as conservative in §2.1.
  • domain assumption The neutral fraction x_HI(z) follows a tanh model (Eq. 3) with parameters z_r and Δz.
    Empirical model from CMB/global-signal analyses; the τ21(z) redshift shape depends on it.
  • ad hoc to paper The spin temperature evolves as T_s(z)=T⋆ 10^{-(2/7)(z-6)} (Eq. 4).
    Toy model for T_s; the paper acknowledges it is simplified, and the text's claimed fiducial behavior is inconsistent with the formula.
  • domain assumption Island model: fully neutral (x_HI=1) rectangular absorption feature of width Δν.
    Scenario for residual HI islands; the T_s>1.73 K limit depends on x_HI=1 within the island.
  • domain assumption The channel covariance matrix is well estimated from the residual image cube (Eq. 7).
    Used to define the likelihood; validated against PyBDSF and Stokes V noise.
  • domain assumption Adiabatic cooling temperature uses z_d=150 (Eq. 8).
    Standard thermal decoupling redshift; sets the comparison value 0.74-0.85 K.

pith-pipeline@v1.3.0-alltime-deepseek · 18772 in / 26303 out tokens · 203398 ms · 2026-08-01T10:02:02.657318+00:00 · methodology

0 comments
read the original abstract

The redshifted 21-cm line is a promising probe of the Epoch of Reionization, during which the first generation of stars ionized the InterGalactic Medium (IGM). We aimed to constrain the IGM neutral Hydrogen fraction and spin temperature via redshifted 21-cm absorption against high-redshift radio sources. We analysed an 18-hour observation of the radio-loud quasar PSO J352.4034-15.3373 ($z = 5.84$) in the 203-222.5 MHz band. We obtained a continuum image with a $0.66$ mJy/beam rms noise and a spectrum with $3.6$ mJy/beam rms noise per 390 kHz-wide channel. We fit the quasar spectrum with a power-law continuum modified by an intervening 21-cm absorption, testing two scenarios: an island model, a residual cold neutral Hydrogen patch surviving at the end of reionization, and a global model, approximating the overall decline of the IGM HI fraction and spin temperature with redshift. Combining archival data spanning 150 MHz to 3.0 GHz with our measurements, we determined a quasar flux density of $86.8 \pm 1.1$ mJy at 200 MHz and a spectral index of $-0.88 \pm 0.02$ across the full frequency range. We found no evidence for the 21-cm absorption from intervening neutral Hydrogen at $5.38<z<5.84$. Assuming the island model, we set a 95% confidence lower limit on the IGM spin temperature of $1.73$ K in residual HI regions. Assuming the global model, we constrained the 21-cm optical depth to $\tau_{21}< 0.02$ (95% C.L.). These results provide constraints on the 21-cm optical depth near the end of reionization, over the $5.38 < z < 5.84$ range, and confirm that the IGM was heated above the adiabatic cooling limit ($\sim 0.8$ K at $z = 5.68$), consistent with theoretical predictions and with 21-cm power spectrum measurements at higher redshifts. Our results also disfavour the presence of extremely cold HI regions at $z < 5.84$ and open the way to future 21-cm absorption from high-redshift sources.

Figures

Figures reproduced from arXiv: 2607.20371 by Andrei Mesinger, Benedetta Ciardi, Chanasorn Kongprachaya, Emilio Ceccotti, Gianni Bernardi, Oleg M. Smirnov.

Figure 1
Figure 1. Figure 1: Simulated spectrum of the hypothetical radio loud source [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Constraints on the flux density at 200 MHz (in mJy) and [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: Reconstructed 21-cm optical depth (τ21) in the global model case as a function of redshift for the hypothetical source z = 8. The red solid line is the input τ21 profile, the black solid line is the best-fit optical depth profile and the light and dark blue shaded regions represent the 68% and 95% C.I., respectively. 5.0 5.5 6.0 6.5 7.0 7.5 8.0 z 0.0 0.2 0.4 0.6 0.8 1.0 x HI [PITH_FULL_IMAGE:figures/full_… view at source ↗
Figure 6
Figure 6. Figure 6: Simulated spectra of radio-loud sources in Tab. 2. The [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 5
Figure 5. Figure 5: Reconstructed HI fraction (xHI) as a function of redshift for the simulated source at z = 8. The red solid line is the input profile used to simulate the absorbed spectrum. The black solid line is the best-fit profile estimated from the simulation. The light and dark blue shaded regions represent the 68% and 95% C.I., respectively. Radio Telescope (uGMRT, proposal code: ddtC007). In particular, we used arc… view at source ↗
Figure 7
Figure 7. Figure 7: Same as Fig. 2, but for each source in the sample of Tab. 2. [PITH_FULL_IMAGE:figures/full_fig_p006_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Constraints on the 21-cm optical depth from simulated [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗
Figure 10
Figure 10. Figure 10: The simulated spectrum of J2329-1520 (blue data points) [PITH_FULL_IMAGE:figures/full_fig_p007_10.png] view at source ↗
Figure 12
Figure 12. Figure 12: Same as Fig. 8, but for simulated observations of J2329- [PITH_FULL_IMAGE:figures/full_fig_p008_12.png] view at source ↗
Figure 15
Figure 15. Figure 15: Continuum image of the J2329-1520 (PSO J352.4034–15.3373) quasar at 212.75 MHz. The black cross marks the pixel with peak flux density. Green contours are drawn at 3σ and 5σ levels, respectively, with the noise rms of σ = 0.66 mJy beam−1 . The image angular resolution is 15.6 ′′ × 11.6 ′′ . 10 3 (MHz) 10 1 10 2 S ( m J y ) [PITH_FULL_IMAGE:figures/full_fig_p009_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: The radio continuum spectrum of J2329-1520 from [PITH_FULL_IMAGE:figures/full_fig_p009_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Spectrum of the J2329-1520 quasar. The frequency [PITH_FULL_IMAGE:figures/full_fig_p010_17.png] view at source ↗
Figure 20
Figure 20. Figure 20: Same as Fig. 8, but for the J2329-1520 observed spectrum. [PITH_FULL_IMAGE:figures/full_fig_p010_20.png] view at source ↗
Figure 22
Figure 22. Figure 22: Same as Fig. 17, but for the [PITH_FULL_IMAGE:figures/full_fig_p011_22.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

47 extracted references · 3 linked inside Pith

  1. [1]

    E., Alexander, P., et al

    Abdurashidova, Z., Aguirre, J. E., Alexander, P., et al. 2022, ApJ, 924, 51 Bañados, E., Carilli, C., Walter, F., et al. 2018, ApJ, 861, L14 Bañados, E., Mazzucchelli, C., Momjian, E., et al. 2021, ApJ, 909, 80 Bañados, E., Momjian, E., Connor, T., et al. 2025, Nature Astronomy, 9, 293

    2022

  2. [2]

    D., Bolton, J

    Becker, G. D., Bolton, J. S., & Lidz, A. 2015, PASA, 32, e045

    2015

  3. [3]

    2020, A&A, 635, L7

    Belladitta, S., Moretti, A., Caccianiga, A., et al. 2020, A&A, 635, L7

    2020

  4. [4]

    J., Mitchell, D

    Bernardi, G., Greenhill, L. J., Mitchell, D. A., et al. 2013, ApJ, 771, 105

    2013

  5. [5]

    Bosman, S. E. I., Davies, F. B., Becker, G. D., et al. 2022, MNRAS, 514, 55

    2022

  6. [6]

    L., Gnedin, N

    Carilli, C. L., Gnedin, N. Y ., & Owen, F. 2002, ApJ, 577, 22 CASA Team, Bean, B., Bhatnagar, S., et al. 2022, PASP, 134, 114501

    2002

  7. [7]

    R., Koopmans, L

    Ceccotti, E., Offringa, A. R., Koopmans, L. V . E., et al. 2025, A&A, 696, A56

    2025

  8. [8]

    2013, MNRAS, 428, 1755

    Ciardi, B., Labropoulos, P., Maselli, A., et al. 2013, MNRAS, 428, 1755

    2013

  9. [9]

    J., Cotton, W

    Condon, J. J., Cotton, W. D., Greisen, E. W., et al. 1998, AJ, 115, 1693

    1998

  10. [10]

    L., van Velzen, S., & Falcke, H

    Coppejans, R., Cseh, D., Williams, W. L., van Velzen, S., & Falcke, H. 2015, MNRAS, 450, 1477

    2015

  11. [11]

    Cornwell, T. J. 2008, arXiv e-prints, arXiv:0806.2228 de Oliveira-Costa, A., Tegmark, M., Gaensler, B. M., et al. 2008, MNRAS, 388, 247 de Vries, W. H., Morganti, R., Röttgering, H. J. A., et al. 2002, AJ, 123, 1784

    Pith/arXiv arXiv 2008

  12. [12]

    J., et al

    Drouart, G., Seymour, N., Galvin, T. J., et al. 2020, PASA, 37, e026

    2020

  13. [13]

    F., & Davies, F

    Eilers, A.-C., Hennawi, J. F., & Davies, F. B. 2018, ApJ, 867, 30

    2018

  14. [14]

    W., Lang, D., & Goodman, J

    Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman, J. 2013, PASP, 125, 306

    2013

  15. [15]

    I., Gabányi, K

    Frey, S., Paragi, Z., Gurvits, L. I., Gabányi, K. É., & Cseh, D. 2011, A&A, 531, L5

    2011

  16. [16]

    P., & Briggs, F

    Furlanetto, S., Oh, S. P., & Briggs, F. 2006, Phys. Rept., 433, 181

    2006

  17. [17]

    R., Oh, S

    Furlanetto, S. R., Oh, S. P., & Briggs, F. H. 2006, Phys. Rep., 433, 181

    2006

  18. [18]

    2024, MNRAS, 530, 191

    Ghara, R., Bag, S., Zaroubi, S., & Majumdar, S. 2024, MNRAS, 530, 191

    2024

  19. [19]

    J., Saxena, A., Intema, H., et al

    Gloudemans, A. J., Saxena, A., Intema, H., et al. 2023, A&A, 678, A128

    2023

  20. [20]

    Gunn, J. E. & Peterson, B. A. 1965, ApJ, 142, 1633 HERA Collaboration, Abdurashidova, Z., Adams, T., et al. 2023, ApJ, 945, 124

    1965

  21. [21]

    R., Hancock, P

    Hurley-Walker, N., Callingham, J. R., Hancock, P. J., et al. 2017, MNRAS, 464, 1146

    2017

  22. [22]

    T., Jagannathan, P., Mooley, K

    Intema, H. T., Jagannathan, P., Mooley, K. P., & Frail, D. A. 2017, A&A, 598, A78

    2017

  23. [23]

    M., Thyagarajan, N., Kumar, A., Kanekar, N., & Bernardi, G

    Keller, P. M., Thyagarajan, N., Kumar, A., Kanekar, N., & Bernardi, G. 2024, MNRAS, 528, 5692

    2024

  24. [24]

    C., Haehnelt, M

    Kulkarni, G., Keating, L. C., Haehnelt, M. G., et al. 2019, MNRAS, 485, L24

    2019

  25. [25]

    2006, MNRAS, 373, 561

    Lewis, A., Weller, J., & Battye, R. 2006, MNRAS, 373, 561

    2006

  26. [26]

    P., et al

    Matsuoka, Y ., Decarli, R., Farina, E. P., et al. 2026, arXiv e-prints, arXiv:2606.10160

    Pith/arXiv arXiv 2026

  27. [27]

    D., Becker, R

    McGreer, I. D., Becker, R. H., Helfand, D. J., & White, R. L. 2006, ApJ, 652, 157

    2006

  28. [28]

    G., Mevius, M., Koopmans, L

    Mertens, F. G., Mevius, M., Koopmans, L. V . E., et al. 2025, A&A, 698, A186

    2025

  29. [29]

    & Rafferty, D

    Mohan, N. & Rafferty, D. 2015, PyBDSF: Python Blob Detection and Source

    2015

  30. [30]

    L., & McGreer, I

    Momjian, E., Carilli, C. L., & McGreer, I. D. 2008, AJ, 136, 344

    2008

  31. [31]

    A., Greig, B., Bowman, J

    Monsalve, R. A., Greig, B., Bowman, J. D., et al. 2018, ApJ, 863, 11

    2018

  32. [32]

    D., Null, D., Trott, C

    Nunhokee, C. D., Null, D., Trott, C. M., et al. 2025, ApJ, 989, 57

    2025

  33. [33]

    G., et al

    Ocvirk, P., Aubert, D., Sorce, J. G., et al. 2020, MNRAS, 496, 4087

    2020

  34. [34]

    R., et al

    Ocvirk, P., Gillet, N., Shapiro, P. R., et al. 2016, MNRAS, 463, 1462

    2016

  35. [35]

    R., McKinley, B., Hurley-Walker, et al

    Offringa, A. R., McKinley, B., Hurley-Walker, et al. 2014, MNRAS, 444, 606

    2014

  36. [36]

    Offringa, A. R. & Smirnov, O. 2017, MNRAS, 471, 301

    2017

  37. [37]

    R., van de Gronde, J

    Offringa, A. R., van de Gronde, J. J., & Roerdink, J. B. T. M. 2012, A&A, 539, A95

    2012

  38. [38]

    Perley, R. A. & Butler, B. J. 2017, ApJS, 230, 7 Planck Collaboration, Adam, R., Aghanim, N., et al. 2016, A&A, 596, A108 Planck Collaboration, Aghanim, N., Akrami, Y ., et al. 2020, A&A, 641, A6

    2017

  39. [39]

    2025, PASA, 42, e049

    Qin, Y ., Mesinger, A., Prelogovi´c, D., et al. 2025, PASA, 42, e049

    2025

  40. [40]

    2020, A&A, 641, A85

    Shao, Y ., Wagg, J., Wang, R., et al. 2020, A&A, 641, A85

    2020

  41. [41]

    2022, A&A, 659, A159

    Shao, Y ., Wagg, J., Wang, R., et al. 2022, A&A, 659, A159

    2022

  42. [42]

    W., Röttgering, H

    Shimwell, T. W., Röttgering, H. J. A., Best, P. N., et al. 2017, A&A, 598, A104

    2017

  43. [43]

    Spinelli, M., Bernardi, G., & Santos, M. G. 2019, MNRAS, 489, 4007

    2019

  44. [44]

    2020, Astrophys

    Thyagarajan, N. 2020, Astrophys. J., 899, 16 Šoltinský, T., Kulkarni, G., Tendulkar, S. P., & Bolton, J. S. 2025, MNRAS, 537, 364

    2020

  45. [45]

    L., van Weeren, R

    Williams, W. L., van Weeren, R. J., Röttgering, H. J. A., et al. 2016, MNRAS, 460, 2385

    2016

  46. [46]

    J., Delorme, P., Reylé, C., et al

    Willott, C. J., Delorme, P., Reylé, C., et al. 2010, AJ, 139, 906

    2010

  47. [47]

    F., Guarneri, F., et al

    Yang, D., Hennawi, J. F., Guarneri, F., et al. 2026, arXiv e-prints, arXiv:2607.03432 Article number, page 12 of 15 C. Kongprachaya et al.: The burgeoning rise of the 21-cm forest Appendix A: Literature measurements for the high redshift quasar sample In this Appendix, we report the literature measurements used to fit the continuum spectrum of the high-re...

    Pith/arXiv arXiv 2026