{"id":"ebd59fc9-6ef3-47b7-957b-3698a98b609f","arxiv_id":"2412.08995","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"A galaxy-catalog-based empirical model reproduces observed radio luminosity functions and source counts, and predicts SKA will find many z=4-6 galaxies invisible to Rubin.","lead":"This paper builds a mock catalog of galaxies and assigns each one a radio brightness from its star formation rate and the chance it hosts a radio-emitting black hole. The simulation matches observed radio source counts and predicts that about half of the distant radio sources the SKA will see are too faint for Rubin's deep optical survey.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The SKA/Rubin prediction rests on an unquantified extrapolation of the Wang et al. AGN probability and RLF to 4<z<6; a sensitivity test freezing the evolution at z=4 should be required.","rationale":"The paper's main contribution is a mock catalog linking radio emission to host galaxy properties, with a concrete prediction for SKA/Rubin synergy. The z<4 validations are strong: differential number counts at 150 MHz, 1.4 GHz, and 3 GHz agree with independent observations, and the SFG RLF matches literature data. The circularity in the AGN RLF validation is real but not fatal, because the assignment convolved the Wang et al. functions with the EGG mass function and AGN fractions; still, it weakens the AGN-specific validation. The most consequential and least supported claim is the z>4 SKA/Rubin fraction, which is the centerpiece of the abstract and Section 4. It depends on equations that the authors explicitly extrapolate (Section 5.1(4)). No uncertainty propagation or alternative model is provided. A concrete sensitivity test—freezing the evolution at z=4—would determine whether the prediction is an artifact of the extrapolation. This does not change the reader's CONDITIONAL verdict: the model is promising and well-validated at lower redshifts, but the high-z prediction should be accompanied by a sensitivity analysis and a data release link. The reader's weakest_assumption identifies the same extrapolation concern, so I agree.","tokens_in":19386,"tokens_out":7436,"duration_ms":77729,"concrete_test":"Re-run the mock catalog with Equations (2) and (5) frozen at their z=4 values for all 4<z<6 (no further redshift evolution) and recompute the fraction of SKA-detected sources with r>27.5 in Figure 6. If the fraction shifts by more than ~10 percentage points, the headline prediction is not robust to the extrapolation and should be reported with an uncertainty band; if it stays within that range, the concern is mitigated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 4.1's headline result—56% of SKA-detected sources at z>4 will be invisible to Rubin's r~27.5—is generated from radio AGNs assigned using Equations (2) and (5), whose redshift evolution is measured only at z<4 and is explicitly extrapolated to 4<z<6 in Section 5.1(4). The high-z AGN RLF is not independently validated: Figure 1 compares against Wang et al. (2024), the same functions used as input, while Figure 2 literature data at z>2 are sparse and mostly at high luminosities. The paper states the extrapolation succeeds in recovering z>4 RLFs, but no quantitative uncertainty or alternative evolution is explored. Because the AGN fraction and luminosity assignment determine what radio sources exist at z~4-6, and because optically faint hosts are preferentially massive, dust-obscured systems, the predicted optically invisible fraction is directly sensitive to the assumed high-z AGN abundance. The lack of a sensitivity analysis means the central SKA/Rubin prediction is not yet robust.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper constructs an empirical model of the extragalactic radio sky by starting from an EGG-generated NIR-selected mock galaxy catalog and assigning radio continuum emission to every galaxy: star-forming galaxies receive 1.4 GHz luminosities from their SFRs through qIR, while radio AGNs are drawn from the Wang et al. (2024) probability functions and assigned luminosities following a pure density evolution RLF. The model also adds Soneira-Peebles clustering for massive galaxies and empirically calibrated sizes for SFGs. The authors validate the resulting AGN and SFG 1.4 GHz RLFs, differential source counts at 150 MHz, 1.4 GHz, and 3 GHz, and the angular two-point correlation function of radio AGNs. They then use the catalog to predict SKA source statistics and LSST cross-matching, finding that roughly half of SKA sources at z>4 will be fainter than r=27.5, and they describe the columns of the released 4 deg^2 mock catalog.","tokens_in":19707,"tokens_out":7756,"duration_ms":85797,"significance":"If the model is correct, it provides a useful physical link between radio flux densities and galaxy host properties, and the released mock catalog can support survey design and multiwavelength follow-up planning. The SFG RLF and the source-count checks use independent data, and the paper is commendably explicit about its limitations, including the extrapolation to z>4, the missing extended radio sources, and the ad hoc 10^10 M_sun host-mass threshold. The main weakness is that the AGN population is generated from the same Wang et al. (2024) functions that are used as the primary AGN validation, so the independent evidence for the high-redshift AGN behavior is sparse; consequently, the headline SKA/LSST prediction requires a sensitivity analysis before it can be regarded as robust.","major_comments":[{"comment":"The calculation of fAGN is not dimensionally consistent as written. In Eq. (2), p(L|M,z) is written as a dimensionless power law, but Eq. (3) integrates it over dL_radio. Since p is proportional to L^-0.77, the integral from L_limit to infinity diverges at the upper end, and even with a finite upper cutoff fAGN would have units of luminosity rather than being a fraction. If p is instead a probability per logarithmic luminosity, Eq. (3) must integrate over dlog L and the normalization in Eq. (2) must be interpreted accordingly. This step controls the total number of radio AGNs in every mass and redshift bin, so it should be fixed and stated explicitly.","section":"§2.2, Eqs. (2)–(3)"},{"comment":"The agreement between the mock AGN RLF and the Wang et al. (2024) data points in Fig. 1 is partly circular: the mock AGNs are generated from the same best-fit functions in Eqs. (2) and (5), so Fig. 1 largely demonstrates that the sampling code recovers its input. The genuinely informative tests are Figs. 2 and 3, which use independent compilations. I request a quantitative comparison in the overlapping luminosity and redshift bins, for example residuals or a chi-square statistic, and a clear separation in the text between internal consistency checks and external validation when summarizing the claim that the AGN RLF is successfully recovered.","section":"§3.1, Fig. 1 vs. §2.2"},{"comment":"The headline prediction that 56% of SKA-detected sources at z>4 are invisible in the LSST r~27.5 survey is driven by radio AGNs generated with Eqs. (2) and (5) extrapolated from z<4, as the text acknowledges, and the high-redshift validation data are sparse. No sensitivity test is provided. I request a robustness check that, at minimum, freezes the AGN probability and RLF evolution at z=4, and ideally also varies the 10^10 M_sun host-mass threshold of §2.2 and the qIR extrapolation for SFGs beyond z~4.5, and reports the resulting range for the claimed 45–56% optically invisible fraction. Without such a test, the main SKA/LSST prediction is not yet demonstrated.","section":"§4.1 and §5.1(4)"}],"minor_comments":[{"comment":"The pivot values M0, L0, and z0 in Eq. (2) are not defined; please state them explicitly.","section":"§2.2, Eq. (2)"},{"comment":"The flux limits, selection criteria, and redshift binning of the literature data used in Figs. 2 and 3 are described in the text but would be easier to assess if summarized in the captions or a table.","section":"§3.1, Figs. 2–3"},{"comment":"The statement that the extrapolations 'succeeded in recovering RLFs of both radio AGNs and SFGs at z>4' is stronger than the sparse z>4 data warrant; please soften it or add quantitative uncertainties.","section":"§5.1(4)"},{"comment":"The LSST r band is approximated with the SDSS r band; the potential filter and k-correction differences should be stated or estimated.","section":"§4.1"},{"comment":"The mock catalog is described as publicly released, but no URL or DOI is given; please add access details.","section":"Table 3"},{"comment":"The z>2 effective-radius fit uses only 19 SFGs and the resulting A and alpha values have large uncertainties; this should be noted wherever the size model is applied.","section":"§2.5, Table 1"},{"comment":"The heading 'Coution' is a typo for 'Caution'; other typos include 'wavelenghths' in the introduction and 'simuations' in §4.2.","section":"§5.2"},{"comment":"The constant C in Eq. (8) is not defined or used; please define it or remove it.","section":"§4.3, Eq. (8)"}],"recommendation":"major_revision","confidential_remarks":"The AGN calibration functions are taken from Wang et al. (2024), which shares two authors with this manuscript. This is not in itself improper, but the manuscript should state more explicitly that Fig. 1 is a consistency check rather than an independent validation. The paper's scope is well matched to a technical astronomy journal, and the released catalog is a useful community resource; the requested sensitivity analysis for the z>4 prediction should be feasible within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look if you build mocks or plan radio-optical survey synergies. The core idea is genuinely new relative to S3 and TRECS: instead of drawing radio sources from RLFs, the authors start from the EGG galaxy catalog and assign radio emission via SFR, qIR, and a radio-AGN probability. Every mock radio source therefore comes with host stellar mass, SFR, and multi-band magnitudes—exactly what you need for survey design and for splitting the ERB by galaxy type. The z<4 validation is convincing: source counts at 150 MHz, 1.4, 3, 10, and 15 GHz match independent data, and the SFG RLF is derived from SFR-M* plus qIR rather than fitted to the RLF.\n\nThe soft spots are real but mostly where you'd expect. The AGN RLF validation is partly circular: mock AGNs are generated using the Wang et al. (2024) probability and RLF shapes (Eqs. 2 and 5), then compared to Wang et al. (2024) data, with two co-authors in common. That doesn't make it worthless—the source counts are independent checks of normalization and spectral handling—but the AGN RLF figure is a consistency check more than a validation. The clustering treatment also includes tuned fractions (15% active, 25% passive in the Soneira-Peebles redistribution) fitted to reproduce the target TPCF; fine as a modeling choice, but not a prediction.\n\nThe bigger concern, and the stress-test note has it right, is the z>4 result. Section 5.1(4) explicitly says Eqs. 2 and 5 are extrapolated to 4<z<6, and the Delvecchio qIR relation is extrapolated past z=4.5. So the headline—56% of SKA sources at z>4 invisible to Rubin's r~27.5—depends directly on the high-z AGN abundance, which is unmeasured. There is no sensitivity test, e.g., freezing evolution at z=4, so we don't know if the fraction is 40% or 70%. This doesn't break the method; it means that specific number is a motivated extrapolation with unquantified uncertainty, not yet a planning-grade constraint.\n\nThe paper is honest about its limitations (extended radio sources, low-mass AGN host threshold, passive galaxy fractions at z>4), the citations look appropriate, and the mock catalog is promised public. Take the z<4 catalog seriously; treat the z>4 fraction with caution; ask for a sensitivity test in revision. I'd send it to a referee—the method is a step forward and the issues are fixable.","headline":"Useful new mock catalog with convincing z<4 validation; the z>4 SKA/Rubin prediction is an explicit extrapolation that needs a sensitivity test.","tokens_in":20272,"tokens_out":3252,"would_cite":true,"duration_ms":28984,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Building the extragalactic radio background from a mock galaxy catalog, this paper predicts that roughly half of the radio sources SKA will detect at z ~ 4–6 will be too faint for Rubin's deepest optical survey.","keywords":["extragalactic radio background","radio luminosity function","mock galaxy catalog","star-forming galaxies","radio AGN","SKA","LSST/Rubin","source counts"],"falsifier":"Measure the 1.4 GHz radio luminosity function of AGNs at 4<z<6 with a deep survey, or measure the fraction of SKA-detected z~4–6 sources with r>27.5 in early SKA/Rubin overlap; a deviation from the extrapolated pure-density-evolution model, or a fraction far from ~50%, would refute the central prediction.","tokens_in":19146,"feed_emoji":"📡","tokens_out":5837,"duration_ms":53633,"temperature":0.7,"pith_summary":"Recent deep radio surveys across 0<z<4 have pinned down how radio emission relates to star formation rate and to the probability of hosting a radio-loud AGN. This paper uses those relations to add radio fluxes to a mock multi-wavelength galaxy catalog, creating a synthetic sky whose 1.4 GHz luminosity functions for both AGN and star-forming galaxies, and whose differential source counts at 150 MHz, 1.4 GHz, and 3 GHz, match observed data. Because the model starts from a galaxy catalog rather than from luminosity functions, every radio source inherits a redshift, stellar mass, and optical/NIR magnitudes. The central forward-looking result is that in the SKA era, roughly half of the radio continuum sources detected at z~4–6 will be fainter than r~27.5 and therefore invisible to Rubin's deepest optical survey.","feed_headline":"Half of SKA's early-universe radio sources would be missed by Rubin","feed_subtitle":"A mock galaxy catalog predicts that about half of SKA's z~4–6 radio sources will have no optical counterpart.","key_machinery":"The carrying object is the EGG mock galaxy catalog (Schreiber et al. 2017), a near-infrared-selected simulated sky containing active and passive galaxies with stellar masses, SFRs, and SEDs down to z~10. Onto it the paper layers two radio channels: star-forming galaxies receive 1.4 GHz luminosities by converting their IR luminosities with the mass- and redshift-dependent qIR calibration of Delvecchio et al. (2021), while radio AGNs are assigned with probability functions p(L1.4GHz|M*, z) from Wang et al. (2024), restricted to hosts above $10^{10}$ M_sun, with luminosities drawn from a pure density evolution luminosity function. A Soneira–Peebles hierarchical clustering algorithm redistributes the massive host galaxies so that the mock radio AGNs reproduce the observed angular two-point correlation function. These components convert ordinary galaxy properties into a realistic radio sky whose multi-wavelength connections can be queried directly.","core_discovery":"The paper's central claim is that the extragalactic radio background can be built from the bottom up: instead of drawing sources from measured radio luminosity functions, it assigns radio emission to every galaxy in the EGG mock catalog based on that galaxy's star formation rate and on a probability, calibrated at z<4, that it hosts a radio AGN of a given 1.4 GHz luminosity. The synthetic catalog then reproduces the observed 1.4 GHz radio luminosity functions of both AGNs and star-forming galaxies, the differential number counts at 150 MHz, 1.4 GHz, and 3 GHz, and the angular clustering of radio AGNs. Since the radio flux is attached to physical galaxies, the catalog directly links radio detection to host properties such as stellar mass, redshift, and photometric magnitudes. Its headline prediction is that 45–56% of the radio sources SKA will detect at z>4 will be too faint at r-band (r~27.5) to be seen by Rubin's LSST survey, making deep radio observations essential for finding dust-obscured early-universe galaxies.","pith_inferences":["The same catalog-based technique could be used to predict X-ray or sub-mm emission from the same mock galaxies, yielding a fully multi-wavelength mock sky for survey planning.","The 'half invisible' prediction would shift if high-redshift radio AGNs are hosted by lower-mass galaxies than assumed; SKA pathfinder data at 4<z<6 could test the extrapolated AGN probabilities within a few years.","If the extrapolation is wrong, the model would still reproduce z<4 observables, meaning the high-z prediction should be read as the main falsifiable output rather than the calibrations."],"forward_implications":["The released 4 deg^2 mock catalog lets survey planners compute, for any radio detection, the expected optical/NIR magnitude and stellar mass of its host.","SKA's 5-hour pointings should detect star-forming galaxies down to ~10^10 M_sun at z~6, extending galaxy formation studies to the early universe.","At z>4, 45–56% of SKA-detected radio sources will lack an LSST r-band (or z-band) counterpart, so SKA will be the primary window onto dust-obscured early galaxies.","Radio AGN and SFG contributions to the extragalactic radio background separate cleanly in the model, but the AGN contribution is sensitive to a single bright source per square degree at 150 MHz."],"supporting_citations":[{"why":"Supplies the EGG mock galaxy catalog with stellar masses, SFRs, and SEDs that the radio assignment builds on.","marker":"Schreiber et al. (2017)"},{"why":"Provides the radio-AGN probability functions and the pure density evolution luminosity function shape used to assign AGNs.","marker":"Wang et al. (2024)"},{"why":"Calibrates qIR as a function of stellar mass and redshift, converting IR luminosity to 1.4 GHz radio luminosity for star-forming galaxies.","marker":"Delvecchio et al. (2021)"},{"why":"Gives the SFR-to-1.4 GHz luminosity relation used to set the lower-luminosity limit for AGN assignment.","marker":"Bell (2003)"},{"why":"The hierarchical clustering algorithm used to reproduce the observed angular clustering of radio AGNs.","marker":"Soneira & Peebles (1978)"},{"why":"Provides the LSST survey depths (r~27.5, z~26.1) used to predict the fraction of SKA-detected sources invisible to Rubin.","marker":"Ivezić et al. (2019)"}],"fun_headline_variants":["Radio model predicts SKA finds galaxies Rubin misses","Bottom-up radio background model: from SFR to 1.4 GHz","Half of SKA's z>4 sources invisible to Rubin's optical survey","New empirical model ties radio emission to galaxy star formation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The high-redshift predictions assume that the radio-AGN probability functions and the luminosity-function shape measured at z<4 continue to hold at 4<z<6 without new data; if those functions change, the 'half invisible' result changes.","fun_headline_variants_meta":{"raw":{"variants":["Radio model predicts SKA finds galaxies Rubin misses","Bottom-up radio background model: from SFR to 1.4 GHz","Half of SKA's z>4 sources invisible to Rubin's optical survey","New empirical model ties radio emission to galaxy star formation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00044,"raw_usage":{"total_tokens":2319,"prompt_tokens":1121,"completion_tokens":1198,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":737,"completion_tokens_details":{"reasoning_tokens":1126}},"tokens_in":737,"tokens_out":1198,"duration_ms":9261,"temperature":1.0,"reasoning_tokens":1126,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T17:21:46.245364+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the 1.4 GHz radio luminosity function of AGNs at 4<z<6 with a deep survey, or measure the fraction of SKA-detected z~4–6 sources with r>27.5 in early SKA/Rubin overlap; a deviation from the extrapolated pure-density-evolution model, or a fraction far from ~50%, would refute the central prediction.","supporting_citations":[{"cited_title":"2017, A&A, 602, A96","cited_arxiv_id":null,"evidence_quote":"Supplies the EGG mock galaxy catalog with stellar masses, SFRs, and SEDs that the radio assignment builds on."},{"cited_title":"Cosmic evolution of radio-excess AGNs in quiescent and star-forming galaxies across $0 < z < 4$","cited_arxiv_id":"2401.04924","evidence_quote":"Provides the radio-AGN probability functions and the pure density evolution luminosity function shape used to assign AGNs."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the SFR-to-1.4 GHz luminosity relation used to set the lower-luminosity limit for AGN assignment."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The hierarchical clustering algorithm used to reproduce the observed angular clustering of radio AGNs."}],"review_version":1}