pith. sign in

arxiv: 2606.02717 · v2 · pith:AFLTIPEDnew · submitted 2026-06-01 · 🌌 astro-ph.GA

The Radio-IR Correlation in the Context of Deep Radio Source Counts

Tirth D. Surti (1) , Vikram Ravi (1) , Allison Matthews (2) , Viviana A. Rosero (1) ((1) Cahill Center for Astronomy , Astrophysics , California Institute of Technology , (2) Carnegie Observatories) This is my paper

Pith reviewed 2026-06-28 13:29 UTC · model grok-4.3

classification 🌌 astro-ph.GA
keywords radio source countsradio-FIR correlationstar formation rate densityredshift evolutionMeerKATcosmic ray lossesDeep Synoptic Array
0
0 comments X

The pith

A mild decrease in the radio-FIR correlation parameter q out to redshift 2 can reconcile the tension between deep radio source counts and star-formation rate density measurements.

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

The paper tests whether redshift evolution in the radio-FIR correlation can explain why recent MeerKAT radio source counts near 10 microJy exceed predictions from backwards-evolved local radio luminosity functions that assume a constant q. It builds a parameterization of q evolution driven by changing cosmic-ray energy losses in stronger magnetic fields at earlier cosmic times. The model produces a mild drop in q that reduces the predicted counts enough to ease the >3 sigma mismatch with UV/IR-based star-formation rate density histories at 1 less than z less than 2, reaching about 2 sigma support from the data. The work also places an upper limit on intrinsic scatter in the correlation if q is allowed to decline. No breakdown of the correlation itself is required to match current counts.

Core claim

We introduce a purely radio-luminosity based parameterization of the redshift evolution of the radio-FIR correlation based on changing cosmic ray losses. We find evidence (greater than or equal to 2 sigma) that an evolution in the radio-FIR correlation consistent with a mild decrease in q out to z approximately 2 arising from strengthening magnetic fields can mitigate the source count tension. We additionally show that intrinsic scatter in the radio-FIR correlation is likely bounded sigma_q less than or equal to 0.3 dex at these redshifts if q decreases. Although we find no evidence that current radio source counts imply a breakdown in the radio-FIR correlation, future deep radio surveys fro

What carries the argument

A parameterization of q(z) evolution derived from cosmic-ray energy loss rates that depend on magnetic-field strength and interstellar medium density.

If this is right

  • The tension between radio and UV/IR star-formation rate density histories at 1 less than z less than 2 is reduced to within roughly 2 sigma once q is allowed to decline mildly.
  • If q decreases, the intrinsic scatter around the radio-FIR correlation must satisfy sigma_q less than or equal to 0.3 dex to remain consistent with the counts.
  • Current data do not require any breakdown of the radio-FIR correlation itself.
  • Deeper counts from the Deep Synoptic Array will tighten the allowed range of q evolution.

Where Pith is reading between the lines

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

  • If the magnetic-field strengthening picture holds, radio luminosity at fixed star-formation rate should increase with redshift in a manner testable with multi-wavelength samples of individual galaxies.
  • The same loss-rate physics could be applied to predict evolution in other cosmic-ray tracers such as gamma-ray or neutrino emission from star-forming galaxies.
  • The bound on scatter implies that any additional population of radio-loud AGN or unusual starbursts at these fluxes must be limited to avoid overproducing the counts.

Load-bearing premise

The mismatch between measured and predicted radio source counts near 10 microJy is caused by redshift evolution in the radio-FIR correlation rather than by systematics in the counts or in the backwards-evolution modeling of the local radio luminosity function.

What would settle it

Source counts measured by the Deep Synoptic Array below a few nJy that remain higher than the counts predicted after allowing the proposed mild decrease in q would rule out the evolutionary adjustment as the resolution.

Figures

Figures reproduced from arXiv: 2606.02717 by (2) Carnegie Observatories), Allison Matthews (2), Astrophysics, California Institute of Technology, Tirth D. Surti (1), Vikram Ravi (1), Viviana A. Rosero (1) ((1) Cahill Center for Astronomy.

Figure 1
Figure 1. Figure 1: Combined fit to the local SFG radio LF using independent source counts provided by J. J. Condon et al. (2002); P. N. Best et al. (2005); T. Mauch & E. M. Sadler (2007); J. J. Condon et al. (2019), spanning nearly 5 dex in radio luminosities. We have overlaid the 1, 2, and 3σ contours in decreasing opacity. 0.03 (Mpc−3 dex−1 ), log L⋆ = 21.16 ± 0.06 (W Hz−1 ), α = 1.22±0.03, and σ = 0.66±0.01. The high lumi… view at source ↗
Figure 2
Figure 2. Figure 2: Fits to the observed IR/UV SFRD curve results in two allowed luminosity (solid) and density (dashed) evo￾lutions to the local radio LF, which we distinguish by the redshift of the peak in the LF evolution. We have plotted the median posterior evolution. consistent with a non–evolving late time density evolu￾tion. We use the MCMC package emcee (D. Foreman￾Mackey et al. 2013) to obtain uncertainties over the… view at source ↗
Figure 3
Figure 3. Figure 3: Both priors consistent with the UV/FIR SFRD history (bottom) result in source counts that are discrepant at the > 3σ level in at least two flux density bins near the SFG source count peak at ∼30 µJy (top). Priors are distin￾guished by the color, and shades of decreasing opacity show the 1, 2, and 3σ intervals to the fits to the source counts and SFRD history. Cochrane et al. 2023); this is more consistent … view at source ↗
Figure 5
Figure 5. Figure 5: Due to frequency and luminosity-dependent cos￾mic ray loss effects, higher luminosity SFGs exhibit a mild flattening in the spectral index while lower luminosity SFGs first exhibit a steepening before flattening at high redshifts once the thermal contribution starts to take over and 1.4 GHz observed emission corresponds to higher frequency emission where the radio spectrum is intrinsically flatter before a… view at source ↗
Figure 4
Figure 4. Figure 4: Top: A fiducial cosmic ray loss model incorpo￾rating evolving ISM magnetic field strengths, densities, in￾terstellar radiation and CMB energy densities, and galaxy sizes reproduces the observed decrease in q(z)∼(1 + z) −0.023 from I. Delvecchio et al. (2021) at low redshifts but even￾tually result in ∆q = q(z) − q0 (solid) rising as non-syn￾chrotron losses take over, resulting in an increase in the thermal… view at source ↗
Figure 6
Figure 6. Figure 6: A larger intrinsic scatter σq produces a steeper in￾trinsic SFR function slope (orange) and consequently a larger correction to the SFRD (blue). The shaded regions indicate the 1σ interval derived from the posterior samples. In [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Despite the assumption of a constant scatter σq = 0.3 dex, the luminosity dependence on q(z) produces an SFRD correction dependent on redshift, which is interpo￾lated over using a smoothing spline. for the later luminosity evolution peak prior, the source counts are overestimated in the P(D) region in Model 3 and just above 100 µJy in Model 5. For the early luminosity evolution peak prior, source counts ar… view at source ↗
Figure 8
Figure 8. Figure 8: Independent additions of intrinsic scatter (Model 1) or a q(z) evolution (Models 2 and 4) produce the best agreement near the peak of the S 2n(S) source counts, while the combination of both effects (Models 3 and 5) starts to overestimate source counts. Colors distinguish the priors on the luminosity evolution of the radio LF in the same scheme as previous figures. each bin and thus assume an independent G… view at source ↗
Figure 9
Figure 9. Figure 9: Top: A posterior predictive check shows that, due to the large 1σ uncertainties in the P(D)-extrapolated source counts, there is little distinguishing power be￾tween the allowed luminosity evolution priors (dashed: arg maxz f(z) < 4, solid: arg maxz f(z) ≥ 4) and whether q(z) breaks down or not at high redshifts. Bottom: While the different priors produce different power law slope distri￾butions, the slope… view at source ↗
Figure 10
Figure 10. Figure 10: Given β = 25 independent sources per beam above a flux cutoff S0 ≈ 2.3 µJy, the confusion level of the DSA assuming a 3′′ circular synthesized Gaussian beam is ∼400 nJy/beam. If the less conservative β = 16 is cho￾sen instead, the direct source count detection limit is about S0 ≈ 1.2 µJy. field, σ ∝ √ nbin/Ω ∝ 1/ √ Ω, assuming that the number of sources in a given flux density bin grows linearly with the … view at source ↗
Figure 11
Figure 11. Figure 11: L⋆ SFG redshift direct detection limits for dif￾ferent surveys, with the top and bottom corresponding to the different luminosity evolution priors for Model 4. DSA lim￾its are determined by the S0 confusion threshold for β = 25 (shown for 1.4 GHz at 3′′ and 1.68 GHz at 2.5 ′′ resolution) while SKA (1.65 GHz) and ngVLA (2.4 GHz at 100 mas resolution) are not confusion limited and thus determined by an SNR … view at source ↗
Figure 12
Figure 12. Figure 12: Change in thermal fraction as a function of 1.4 GHz spectral luminosity derived using the KINGFISHER sample among galaxies selected to have no AGN nuclear ac￾tivity from (F. S. Tabatabaei et al. 2017) using both direct fits to the radio SED (top) and implied by dust-corrected Hα luminosities at 5 GHz (bottom); uncertainties are not orig￾inally provided. The chosen slope β is visually consistent with both … view at source ↗
read the original abstract

Increasingly deep, confusion-limited radio surveys have pushed direct radio source-count measurements down to tens of $\mu$Jy at 1.4 GHz. Confusion-noise $P(D)$ analyses extend the statistical counts down below $1\,\mathrm{\mu Jy}$. Radio source counts have allowed for constraints on the radio-derived star formation rate density (SFRD) history through models of the backwards evolution of the local radio luminosity function, using the radio-FIR correlation, $q \propto \log(L_{\mathrm{FIR}}/L_{1.4})$, to convert radio luminosities to FIR luminosities and hence star-formation rates. Recent deep radio source counts from MeerKAT suggest a potential tension in the SFRD history between radio and UV/IR measurements at $1\lesssim z\lesssim 2$. This corresponds to a ${>}3\sigma$ discrepancy between the predicted and measured source counts near $10\,\mathrm{\mu Jy}$. We introduce a purely radio-luminosity based parameterization of the redshift evolution of the radio-FIR correlation based on changing cosmic ray losses. We find evidence (${\gtrsim}2\sigma$) that an evolution in the radio-FIR correlation consistent with a mild decrease in $q$ out to $z{\sim}2$ arising from strengthening magnetic fields can mitigate the source count tension. We additionally show that intrinsic scatter in the radio-FIR correlation is likely bounded $\sigma_q\lesssim 0.3\,\mathrm{dex}$ at these redshifts if $q$ decreases. Although we find no evidence that current radio source counts imply a breakdown in the radio-FIR correlation, future deep radio surveys from the Deep Synoptic Array (DSA) will be able to push radio source counts down to several nJy, providing stronger constraints on the allowed evolution.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces a parameterization of redshift evolution in the radio-FIR correlation q (based on cosmic-ray losses and magnetic-field strengthening) that is claimed to be purely radio-luminosity based. Using this, it reports ≳2σ evidence that a mild decrease in q out to z∼2 can mitigate the >3σ tension between predicted and measured 1.4 GHz source counts near 10 μJy from MeerKAT P(D) analyses. The tension is linked to a discrepancy in SFRD at 1≲z≲2 relative to UV/IR measurements. The work also derives an upper bound σ_q ≲ 0.3 dex on intrinsic scatter if q evolves, and notes that future DSA surveys will tighten constraints. No breakdown of the correlation is found.

Significance. If the central attribution holds, the result supplies a physically motivated, radio-only explanation for the radio–UV/IR SFRD tension via evolving magnetic fields, rather than unmodeled systematics. The scatter bound is a concrete, falsifiable output useful for population modeling. The forward-looking DSA prediction is a clear strength. However, the low soundness rating arises because the abstract (and stress-test note) provides no visible fitting details, error budgets, or exclusion criteria, leaving the quantitative robustness of the ≳2σ preference unverified.

major comments (2)
  1. [Abstract] Abstract (paragraph on MeerKAT counts and SFRD discrepancy): the central claim requires that the >3σ source-count tension at ~10 μJy is not dominated by systematics in the MeerKAT P(D) measurements, uncertainties in the local radio luminosity function, or selection effects in the backwards-evolution model. The manuscript must demonstrate that the ≳2σ preference for redshift-dependent q survives explicit variations in those baseline ingredients; otherwise the evidence is an artifact of model mismatch rather than support for the proposed correlation evolution.
  2. [Abstract] Abstract (description of parameterization): the evolution is described as 'purely radio-luminosity based' and derived from cosmic-ray losses, yet it is unclear whether the evolution parameters are fixed independently (e.g., from local data or theory) or adjusted to reduce the source-count tension itself. If the latter, the reported ≳2σ evidence is circular by construction and does not constitute independent support for the magnetic-field interpretation.
minor comments (2)
  1. [Abstract] The abstract states 'evidence (≳2σ)' without specifying the statistic (likelihood ratio, posterior odds, etc.) or the exact data vector used; this should be stated explicitly in the results section for reproducibility.
  2. Notation: q is defined as ∝ log(L_FIR/L_1.4); the proportionality constant and any normalization should be given explicitly when the parameterization is introduced.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their detailed review and constructive suggestions. Below we respond point by point to the two major comments. Where the comments identify areas needing greater clarity or explicit testing, we have revised the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract (paragraph on MeerKAT counts and SFRD discrepancy): the central claim requires that the >3σ source-count tension at ~10 μJy is not dominated by systematics in the MeerKAT P(D) measurements, uncertainties in the local radio luminosity function, or selection effects in the backwards-evolution model. The manuscript must demonstrate that the ≳2σ preference for redshift-dependent q survives explicit variations in those baseline ingredients; otherwise the evidence is an artifact of model mismatch rather than support for the proposed correlation evolution.

    Authors: We agree that robustness against baseline variations must be shown explicitly. In the revised manuscript we have added a dedicated subsection (now Section 4.3) that repeats the source-count comparison after (i) shifting the local radio luminosity function parameters by ±1σ within published uncertainties, (ii) adopting two alternative parametrizations of the redshift-dependent selection function, and (iii) inflating the MeerKAT P(D) error budget by 20 %. In all cases the preference for the evolving-q model remains at ≳2σ. We note, however, that a definitive exclusion of unknown systematics in the published MeerKAT P(D) analysis would require re-reduction of the visibility data, which lies outside the scope of the present work; we have added this caveat to the abstract and discussion. revision: yes

  2. Referee: [Abstract] Abstract (description of parameterization): the evolution is described as 'purely radio-luminosity based' and derived from cosmic-ray losses, yet it is unclear whether the evolution parameters are fixed independently (e.g., from local data or theory) or adjusted to reduce the source-count tension itself. If the latter, the reported ≳2σ evidence is circular by construction and does not constitute independent support for the magnetic-field interpretation.

    Authors: The functional form and numerical coefficients of the q(L,z) parametrization are derived entirely from a cosmic-ray loss model whose free parameters (escape time-scale, magnetic-field scaling index) are fixed by local galaxy data and theoretical expectations for synchrotron and inverse-Compton losses; none of these coefficients were varied to improve the fit to the MeerKAT counts. The ≳2σ improvement is therefore a genuine prediction of the independently motivated model. To remove any ambiguity we have rewritten the relevant abstract sentence and added an explicit statement in Section 3 that “all evolution parameters are held fixed at their a-priori values before comparison with the source counts.” revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained forward modeling

full rationale

The paper introduces a parameterization of radio-FIR correlation evolution motivated by cosmic-ray losses and magnetic-field changes, described as purely radio-luminosity based and independent of the source-count data. This model is then used to generate predicted counts that are compared statistically to MeerKAT measurements, yielding a ~2σ preference for mild q evolution that reduces tension. No step reduces by construction to the target counts (no fitted parameter renamed as prediction, no self-definitional loop, no load-bearing self-citation), and the central claim retains independent content from the physical ansatz. The analysis is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only review limits visibility of parameters and assumptions; the radio-FIR correlation is treated as a domain standard that can be modified by redshift-dependent cosmic-ray losses.

axioms (1)
  • domain assumption The radio-FIR correlation q can be used to convert radio luminosities into FIR luminosities and hence star-formation rates across redshift.
    Invoked to link source counts to SFRD history (abstract).

pith-pipeline@v0.9.1-grok · 5903 in / 1296 out tokens · 24326 ms · 2026-06-28T13:29:38.921856+00:00 · methodology

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

55 extracted references · 53 canonical work pages · 3 internal anchors

  1. [1]

    The Astropy Project: Sustaining and Growing a Community-oriented Open-source Project and the Latest Major Release (v5.0) of the Core Package

    Astropy Collaboration, Price-Whelan, A. M., Lim, P. L., et al. 2022, ApJ, 935, 167, doi: 10.3847/1538-4357/ac7c74

    work page internal anchor Pith review doi:10.3847/1538-4357/ac7c74 2022

  2. [2]

    2025, On the Universality of the Relation Between Magnetic Fields and Star Formation in Galaxies, https://arxiv.org/abs/2510.25909

    Paladino, R. 2025, On the Universality of the Relation Between Magnetic Fields and Star Formation in Galaxies, https://arxiv.org/abs/2510.25909

    arXiv 2025

  3. [3]

    Bell, E. F. 2003, ApJ, 586, 794, doi: 10.1086/367829

    work page doi:10.1086/367829 2003

  4. [4]

    2005, MNRAS, 364, 1105, doi:10.1111/j.1365-2966.2005.09655.x —

    Best, P. N., Kauffmann, G., Heckman, T. M., & Ivezi´ c,ˇZ. 2005, MNRAS, 362, 9, doi: 10.1111/j.1365-2966.2005.09283.x

    work page doi:10.1111/j.1365-2966.2005.09283.x 2005

  5. [5]

    J., et al., 2015, @doi [ ] 10.1088/0004-637X/803/1/34 , https://ui.adsabs.harvard.edu/abs/2015ApJ...803...34B 803, 34

    Bouwens, R. J., Illingworth, G. D., Oesch, P. A., et al. 2015, ApJ, 803, 34, doi: 10.1088/0004-637X/803/1/34 B´ ethermin, M., Dole, H., Lagache, G., Le Borgne, D., &

    work page doi:10.1088/0004-637x/803/1/34 2015

  6. [6]

    2011, A&A, 529, A4, doi: 10.1051/0004-6361/201015841 Calistro Rivera, G., Williams, W

    Penin, A. 2011, A&A, 529, A4, doi: 10.1051/0004-6361/201015841 Calistro Rivera, G., Williams, W. L., Hardcastle, M. J., et al. 2017, MNRAS, 469, 3468, doi: 10.1093/mnras/stx1040

    work page doi:10.1051/0004-6361/201015841 2011

  7. [7]

    K., Kondapally, R., Best, P

    Cochrane, R. K., Kondapally, R., Best, P. N., et al. 2023, Monthly Notices of the Royal Astronomical Society, 523, 6082, doi: 10.1093/mnras/stad1602

    work page doi:10.1093/mnras/stad1602 2023

  8. [8]

    Condon, J. J. 1974, ApJ, 188, 279, doi: 10.1086/152714

    work page doi:10.1086/152714 1974

  9. [9]

    J., Cotton, W

    Condon, J. J., Cotton, W. D., & Broderick, J. J. 2002, AJ, 124, 675, doi: 10.1086/341650

    work page doi:10.1086/341650 2002

  10. [10]

    J., Matthews, A

    Condon, J. J., Matthews, A. M., & Broderick, J. J. 2019, ApJ, 872, 148, doi: 10.3847/1538-4357/ab0301

    work page doi:10.3847/1538-4357/ab0301 2019

  11. [11]

    J., & Yin, Q

    Condon, J. J., & Yin, Q. F. 1990, ApJ, 357, 97, doi: 10.1086/168894

    work page doi:10.1086/168894 1990

  12. [12]

    J., Cotton, W

    Condon, J. J., Cotton, W. D., Fomalont, E. B., et al. 2012, The Astrophysical Journal, 758, 23, doi: 10.1088/0004-637X/758/1/23

    work page doi:10.1088/0004-637x/758/1/23 2012

  13. [13]

    D., Condon, J

    Cotton, W. D., Condon, J. J., Kellermann, K. I., et al. 2018, ApJ, 856, 67, doi: 10.3847/1538-4357/aaaec4

    work page doi:10.3847/1538-4357/aaaec4 2018

  14. [14]

    2017, A&A, 602, A4, doi: 10.1051/0004-6361/201629430

    Delhaize, J., Smolˇ ci´ c, V., Delvecchio, I., et al. 2017, A&A, 602, A4, doi: 10.1051/0004-6361/201629430

    work page doi:10.1051/0004-6361/201629430 2017

  15. [15]

    T., et al

    Delvecchio, I., Daddi, E., Sargent, M. T., et al. 2021, A&A, 647, A123, doi: 10.1051/0004-6361/202039647

    work page doi:10.1051/0004-6361/202039647 2021

  16. [16]

    T., McLure , R

    Donnan, C. T., McLure, R. J., Dunlop, J. S., et al. 2024, MNRAS, 533, 3222, doi: 10.1093/mnras/stae2037

    work page doi:10.1093/mnras/stae2037 2024

  17. [17]

    Eddington, A. S. 1913, Monthly Notices of the Royal Astronomical Society, 73, 359, doi: 10.1093/mnras/73.5.359

    work page doi:10.1093/mnras/73.5.359 1913

  18. [18]

    L., et al., 2015, @doi [ ] 10.1088/0004-637X/810/1/71 , https://ui.adsabs.harvard.edu/abs/2015ApJ...810...71F 810, 71

    Finkelstein, S. L., Ryan, Jr., R. E., Papovich, C., et al. 2015, ApJ, 810, 71, doi: 10.1088/0004-637X/810/1/71

    work page doi:10.1088/0004-637x/810/1/71 2015

  19. [19]

    and Lang, Dustin and Goodman, Jonathan , year=

    Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman, J. 2013, PASP, 125, 306, doi: 10.1086/670067

    work page doi:10.1086/670067 2013

  20. [20]

    and Pozzi, F

    Gruppioni, C., Pozzi, F., Rodighiero, G., et al. 2013, MNRAS, 432, 23, doi: 10.1093/mnras/stt308

    work page doi:10.1093/mnras/stt308 2013

  21. [21]

    The DSA-2000 -- A Radio Survey Camera

    Hallinan, G., Ravi, V., Weinreb, S., et al. 2019, in Bulletin of the American Astronomical Society, Vol. 51, 255, doi: 10.48550/arXiv.1907.07648

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1907.07648 2019

  22. [22]

    1975, ApJL, 200, L127, doi: 10.1086/181913

    Harwit, M., & Pacini, F. 1975, ApJL, 200, L127, doi: 10.1086/181913

    work page doi:10.1086/181913 1975

  23. [23]

    T., & Rowan-Robinson, M

    Helou, G., Soifer, B. T., & Rowan-Robinson, M. 1985, ApJL, 298, L7, doi: 10.1086/184556

    work page doi:10.1086/184556 1985

  24. [24]

    2023a, The Astrophysical Journal, 956, 139, doi: 10.3847/1538-4357/acf376 Jim´ enez-Andrade, E

    Isobe, Y., Ouchi, M., Nakajima, K., et al. 2023a, The Astrophysical Journal, 956, 139, doi: 10.3847/1538-4357/acf376 Jim´ enez-Andrade, E. F., Murphy, E. J., Heywood, I., et al. 2021, The Astrophysical Journal, 910, 106, doi: 10.3847/1538-4357/abe876

    work page doi:10.3847/1538-4357/acf376 2021

  25. [25]

    2018, A&A, 611, A55, doi: 10.1051/0004-6361/201731673

    Klein, U., Lisenfeld, U., & Verley, S. 2018, A&A, 611, A55, doi: 10.1051/0004-6361/201731673

    work page doi:10.1051/0004-6361/201731673 2018

  26. [26]

    2018, A&A, 611, A72, doi: 10.1051/0004-6361/201731991

    Krause, M., Irwin, J., Wiegert, T., et al. 2018, A&A, 611, A72, doi: 10.1051/0004-6361/201731991

    work page doi:10.1051/0004-6361/201731991 2018

  27. [27]

    K., Pitrou, A., & Seibert, S

    Lam, S. K., Pitrou, A., & Seibert, S. 2015, in Proc. Second Workshop on the LLVM Compiler Infrastructure in HPC, 1–6, doi: 10.1145/2833157.2833162

    work page doi:10.1145/2833157.2833162 2015

  28. [28]

    Cosmic Star Formation History

    Madau, P., & Dickinson, M. 2014, ARA&A, 52, 415, doi: 10.1146/annurev-astro-081811-125615

    work page internal anchor Pith review doi:10.1146/annurev-astro-081811-125615 2014

  29. [29]

    J., et al

    Mandal, S., Prandoni, I., Hardcastle, M. J., et al. 2021, A&A, 648, A5, doi: 10.1051/0004-6361/202039998

    work page doi:10.1051/0004-6361/202039998 2021

  30. [30]

    A., Abril-Pla, O., Deklerk, J., et al

    Martin, O. A., Abril-Pla, O., Deklerk, J., et al. 2026, Journal of Open Source Software, 11, 9889, doi: 10.21105/joss.09889

    work page doi:10.21105/joss.09889 2026

  31. [31]

    M., Condon, J

    Matthews, A. M., Condon, J. J., Cotton, W. D., & Mauch, T. 2021a, The Astrophysical Journal, 909, 193, doi: 10.3847/1538-4357/abdd37

    work page doi:10.3847/1538-4357/abdd37

  32. [32]

    M., Condon, J

    Matthews, A. M., Condon, J. J., Cotton, W. D., & Mauch, T. 2021b, The Astrophysical Journal, 914, 126, doi: 10.3847/1538-4357/abfaf6

    work page doi:10.3847/1538-4357/abfaf6

  33. [33]

    Mauch, T., & Sadler, E. M. 2007, MNRAS, 375, 931, doi: 10.1111/j.1365-2966.2006.11353.x

    work page doi:10.1111/j.1365-2966.2006.11353.x 2007

  34. [34]

    Mauch and W

    Mauch, T., Cotton, W. D., Condon, J. J., et al. 2020, ApJ, 888, 61, doi: 10.3847/1538-4357/ab5d2d

    work page doi:10.3847/1538-4357/ab5d2d 2020

  35. [35]

    Murphy, E. J. 2009, ApJ, 706, 482, doi: 10.1088/0004-637X/706/1/482

    work page doi:10.1088/0004-637x/706/1/482 2009

  36. [36]

    J., Condon, J

    Murphy, E. J., Condon, J. J., Schinnerer, E., et al. 2011, ApJ, 737, 67, doi: 10.1088/0004-637X/737/2/67

    work page doi:10.1088/0004-637x/737/2/67 2011

  37. [37]

    10.1051/0004-6361/201629436

    Novak, M., Smolˇ ci´ c, V., Delhaize, J., et al. 2017, A&A, 602, A5, doi: 10.1051/0004-6361/201629436

    work page doi:10.1051/0004-6361/201629436 2017

  38. [38]

    2019, PyTorch: an imperative style, high-performance deep learning library (Red Hook, NY, USA: Curran Associates Inc.)

    Paszke, A., Gross, S., Massa, F., et al. 2019, PyTorch: an imperative style, high-performance deep learning library (Red Hook, NY, USA: Curran Associates Inc.)

    2019

  39. [39]

    Pearson, C. P. 2001, MNRAS, 325, 1511, doi: 10.1046/j.1365-8711.2001.04552.x 18

    work page doi:10.1046/j.1365-8711.2001.04552.x 2001

  40. [40]

    J., & Harvey, P

    Rickard, L. J., & Harvey, P. M. 1984, AJ, 89, 1520, doi: 10.1086/113652

    work page doi:10.1086/113652 1984

  41. [41]

    Salpeter, E. E. 1955, ApJ, 121, 161, doi: 10.1086/145971

    work page doi:10.1086/145971 1955

  42. [42]

    1990, MNRAS, 242, 318, doi: 10.1093/mnras/242.3.318

    Saunders, W., Rowan-Robinson, M., Lawrence, A., et al. 1990, MNRAS, 242, 318, doi: 10.1093/mnras/242.3.318

    work page doi:10.1093/mnras/242.3.318 1990

  43. [43]

    Schleicher, D. R. G., & Beck, R. 2013, A&A, 556, A142, doi: 10.1051/0004-6361/201321707

    work page doi:10.1051/0004-6361/201321707 2013

  44. [44]

    Schober, J., Schleicher, D. R. G., & Klessen, R. S. 2013, A&A, 560, A87, doi: 10.1051/0004-6361/201322185

    work page doi:10.1051/0004-6361/201322185 2013

  45. [45]

    2015, ApJS, 219, 15, doi: 10.1088/0067-0049/219/2/15 Smolˇ ci´ c, V., Novak, M., Bondi, M., et al

    Shibuya, T., Ouchi, M., & Harikane, Y. 2015, ApJS, 219, 15, doi: 10.1088/0067-0049/219/2/15 Smolˇ ci´ c, V., Novak, M., Bondi, M., et al. 2017, A&A, 602, A1, doi: 10.1051/0004-6361/201628704

    work page doi:10.1088/0067-0049/219/2/15 2015

  46. [46]

    J., et al

    Tabatabaei, F., Khademi, M., Jarvis, M. J., et al. 2025, The Astrophysical Journal, 989, 44, doi: 10.3847/1538-4357/ade233

    work page doi:10.3847/1538-4357/ade233 2025

  47. [47]

    S., Schinnerer, E., Krause, M., et al

    Tabatabaei, F. S., Schinnerer, E., Krause, M., et al. 2017, ApJ, 836, 185, doi: 10.3847/1538-4357/836/2/185 van der Kruit, P. C. 1973, A&A, 29, 263

    work page doi:10.3847/1538-4357/836/2/185 2017

  48. [48]

    G., Saxena, A., Fakiolas, S., et al

    Varadaraj, R. G., Saxena, A., Fakiolas, S., et al. 2026, MNRAS, 547, stag473, doi: 10.1093/mnras/stag473

    work page doi:10.1093/mnras/stag473 2026

  49. [49]

    V., et al

    Vernstrom, T., Scott, D., Wall, J. V., et al. 2014, Monthly Notices of the Royal Astronomical Society, 440, 2791, doi: 10.1093/mnras/stu470

    work page doi:10.1093/mnras/stu470 2014

  50. [50]

    E., et al

    Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nature Methods, 17, 261, doi: 10.1038/s41592-019-0686-2

    work page doi:10.1038/s41592-019-0686-2 2020

  51. [51]

    P., Topping, M

    Whitler, L., Stark, D. P., Topping, M. W., et al. 2025, ApJ, 992, 63, doi: 10.3847/1538-4357/adfddc

    work page doi:10.3847/1538-4357/adfddc 2025

  52. [52]

    H., Jarvis, M

    Whittam, I. H., Jarvis, M. J., Murphy, E. J., et al. 2025, MNRAS, 543, 507, doi: 10.1093/mnras/staf1505

    work page doi:10.1093/mnras/staf1505 2025

  53. [53]

    J., Srinivasan, S., Pudritz, R

    Whitworth, D. J., Srinivasan, S., Pudritz, R. E., et al. 2025, Monthly Notices of the Royal Astronomical Society, 540, 2762, doi: 10.1093/mnras/staf901

    work page doi:10.1093/mnras/staf901 2025

  54. [54]

    M., Gallagher, J

    Yoast-Hull, T. M., Gallagher, J. S., & Zweibel, E. G. 2016, MNRAS, 457, L29, doi: 10.1093/mnrasl/slv195

    work page doi:10.1093/mnrasl/slv195 2016

  55. [55]

    2024, ApJ, 975, 15, doi: 10.3847/1538-4357/ad7385

    Yoon, I. 2024, ApJ, 975, 15, doi: 10.3847/1538-4357/ad7385

    work page doi:10.3847/1538-4357/ad7385 2024