{"id":"c2e90dc6-444c-476e-82ba-a42619e018c7","arxiv_id":"2412.04211","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"BEACON's first 19 pure-parallel fields yield 129 galaxy candidates at z>7, 11 at z>10, and zero at z>13, with number densities consistent with earlier JWST and HST surveys.","lead":"The BEACON survey uses spare JWST observing time to image about 100 independent patches of sky, and the first 19 fields already contain 129 photometric galaxy candidates at redshift above 7. The result matters because it offers a cosmic-variance-resistant census of the earliest galaxies, and finds no candidates at redshift beyond 13 in a large volume.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The zero-count at z>13 and its cosmic-variance interpretation rely on a completeness simulation whose z>12 SED pool is a single undifferentiated JAGUAR bin; if real high-z SEDs differ (e.g., higher LyC escape), effective volumes and upper limits shift.","rationale":"The reader's verdict is CONDITIONAL, with the completeness simulation identified as the weakest assumption. I agree with that assessment and sharpen it: the most load-bearing element is the SED prior at z>12, not the size relation per se, because the paper explicitly says the z>12 JAGUAR spectra are combined into one pool. This means the simulation has effectively zero constraining power on the spectral diversity of the very population whose absence it is used to report. A higher LyC escape fraction is a plausible physical alternative (e.g., as inferred for some z~3–6 LBGs and discussed for z>9 candidates), and it would directly break the F150W dropout criterion. The size distribution also matters for aperture photometry, but the SED issue is more specific to the z>13 window. The paper's honest admission that the z>10 selection may be too conservative is an independent signal that incompleteness is not fully controlled; the authors attribute the discrepancy to cosmic variance, but the validation against 12/13 spec-z sources is dominated by z=10–13, not the z>13 range where no such validation is possible. The proposed test directly perturbs the simulation inputs and asks whether the effective volumes change; this is a standard robustness check for LF surveys and is feasible with the released simulation code. If the volumes are stable to the perturbations, the concern is retired; if not, the CONDITIONAL verdict should be maintained until a dedicated z>13 validation (e.g., recovery of JADES-GS-z14-0 in a mock with BEACON depths, or a larger spec-z sample) is performed.","tokens_in":35490,"tokens_out":9451,"duration_ms":93704,"concrete_test":"Rerun the GLACiAR2 completeness simulation for the F150W-dropout selection with two perturbed input sets: (a) SEDs augmented with high Lyman-continuum escape fraction templates (f_esc ~ 0.2–0.5) in addition to the z>12 JAGUAR pool, and (b) half-light radii drawn at 0.5x and 2x the Morishita et al. (2024b) relation. If the recovered effective volumes in the z>13 window (Table 2, F150W rows) change by more than ~30% at M_UV = −21 or −22, the zero-count upper limits are not robust enough to support the cosmic-variance interpretation. As a secondary check, compute the Poisson probability of the observed zero counts from the Table 2 volumes using the Robertson et al. (2024) LF; if P(0) > 5%, the 'no z>13 galaxies' result is statistically consistent with that LF and should not be over-interpreted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central null result—zero F150W-dropout candidates at z>13 and the resulting upper limits in Table 2—is only as strong as the completeness simulation of Sec. 5.3. For z>12, the simulation draws SEDs from a single combined pool of all JAGUAR spectra at z>12 (Sec. 5.3: 'we combine all z>12 JAGUAR spectra into one spectral pool'). This is a deliberately broad but uncalibrated prior in exactly the redshift range where the headline claim is made. If real z>13 galaxies have a higher Lyman-continuum escape fraction, the F150W non-detection criterion (S/N150<2) would fail for otherwise real candidates, and the survey would miss them; if they are more compact than the adopted Morishita et al. (2024b) M_UV-size relation, the fixed 0.16'' aperture photometry would detect a smaller fraction of their light, again lowering completeness. Both effects would reduce the true recovery fraction, making the effective volumes in Table 2 overestimates and the derived upper limits artificially low. The authors' own remark in Sec. 5.2 that the z>10 selection 'seems to be too conservative'—based on the absence of luminous z>10 galaxies relative to legacy surveys—shows that selection incompleteness is a real risk; the same risk has not been excluded at z>13. The recovery of 12/13 literature spec-z>10 sources is encouraging but tests only the bright end and mostly z=10–13, not the z>13 regime.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces BEACON, a JWST Cycle 2 NIRCam pure-parallel survey designed to find z>7 galaxies over about 100 independent sightlines. Using the first 19 fields covering roughly 180 arcmin^2, the authors reduce NIRCam images in six to eight filters, apply Lyman-break dropout plus photometric-redshift selection, and identify 129 z>7 candidates (118 F090W dropouts, 11 F115W dropouts, and zero F150W dropouts). They estimate number densities at 7<z<13 using field-dependent effective volumes from completeness simulations, finding densities overall consistent with previous surveys. The headline result is that no z>13 candidates are found despite a volume of about 10^5 Mpc^3, which the authors interpret as indicating that the bright z>13 sources reported in some legacy fields may be enhanced by cosmic variance. The paper also presents public data products and several ancillary science cases.","tokens_in":35799,"tokens_out":11108,"duration_ms":123159,"significance":"The survey design is well suited to its goal: many independent sightlines reduce cosmic variance relative to contiguous legacy fields, and the initial dataset is a useful community resource. The photometric selection is described carefully and is externally validated against 12 of 13 spectroscopically confirmed z>10 sources, the number densities use Gehrels small-number uncertainties, and the completeness simulation accounts for field-to-field depth variations. If the zero z>13 result is robust, it is an important complement to the small number of legacy fields. The main caveat is that the z>13 upper limits inherit completeness corrections from simulated SEDs and sizes that are not externally calibrated in exactly the redshift range of the headline claim; the authors' own remark in Sec. 5.2 that the z>10 selection 'seems to be too conservative' makes this caveat relevant. With additional robustness tests, the paper would provide a solid quantitative basis for the cosmic-variance interpretation.","major_comments":[{"comment":"The definition of P(Mrecov, z) is internally inconsistent. The text says the numerator counts simulated galaxies 'recovered to have UV magnitude in the bin Mrecov (regardless of their intrinsic UV magnitude)', while the denominator is 'the number of injected galaxies at redshift z with intrinsic UV magnitude equal to Mrecov'. If the numerator includes sources that scatter in from other magnitude bins but the denominator includes only sources with intrinsic magnitude Mrecov, P can exceed unity and the effective volumes in Table 2 are not a standard completeness-weighted volume. Please clarify the intended definition or correct the formula, and state explicitly whether the implementation matches the original Leethochawalit et al. (2023) convention, since Table 2 is load-bearing for both the number densities and the z>13 upper limits.","section":"Sec. 5.3, Eq. (1)"},{"comment":"The z>13 null result rests on a completeness simulation whose redshift range of interest uses a single pooled JAGUAR z>12 spectral bin and sizes drawn from the Morishita et al. (2024b) M_UV-size relation. The 12/13 recovery test in Sec. 5.2 is encouraging, but it validates the selection mostly at z=10-13 and at the bright end, not at z>13; the authors themselves note in Sec. 5.2 that the z>10 selection 'seems to be too conservative'. I request robustness tests of the effective volumes and upper limits against (i) alternative SED shapes, including higher Lyman-continuum escape fractions and bluer or redder UV slopes, and (ii) sizes bracketing the adopted size-luminosity relation by factors of two. If the effective volume in the F150W-dropout rows of Table 2 changes substantially under these bracketing assumptions, the abstract's 'no galaxy candidates at z>13' statement should be correspondingly qualified.","section":"Sec. 5.3 and Table 2"},{"comment":"The interpretation that the zero z>13 count 'may suggest a significant impact from cosmic variance' should be supported by a quantitative statement of the expected number of z>13 sources under the Robertson et al. (2024) LF that contains GS-z14-0, given the BEACON effective volumes. As written, the reader cannot tell whether the absence is a 1-sigma, 2-sigma, or 3-sigma tension with that LF, and the paper's own comparison says the upper limits are consistent with most previous studies. Adding the expected counts and the corresponding Poisson confidence level would make the cosmic-variance claim easier to evaluate.","section":"Sec. 5.2 and Table 2"},{"comment":"For the F150W-dropout selection, the stated redshift window is 13<z<18 but the photometric-redshift threshold is zset=10, so the criterion p(z>10)>0.8 is weaker than p(z>13)>0.8. Please confirm that the F150W non-detection requirement effectively enforces z>13, or change zset to 13 so that the selected sample and the completeness volume are both defined over the same redshift range.","section":"Sec. 4.4"},{"comment":"The F090W-dropout window is quoted as 7.3<z<9.7 with zset=6, while several entries in Table 3 have photometric redshifts near z~7.0 and the paper describes the sample as z>7. Please clarify the adopted lower boundary of the F090W-dropout selection and why some candidates fall below 7.3.","section":"Sec. 4.4 and Table 3"}],"minor_comments":[{"comment":"Please state the total number of injected galaxies per magnitude-redshift bin (the current text says 1200 galaxies are injected 'into each magnitude-redshift bin' but the exact binning of the injection is not fully specified).","section":"Sec. 5.3"},{"comment":"The text says that magnitude bins without sources are shown as 2-sigma upper limits, but the table note says the uncertainties are calculated as in Gehrels (1986), which is usually quoted at 1 sigma. Please specify the confidence level of the upper limits consistently in the table, figure, and text.","section":"Table 2 and Fig. 8"},{"comment":"The word 'complimentary' should be 'complementary' in the abstract and in Sec. 6.","section":"Abstract and Sec. 6"},{"comment":"Adding the expected BEACON count under the Robertson et al. (2024) LF in the z~15 panel would help the reader see whether the zero count is actually in tension with that LF.","section":"Fig. 8"},{"comment":"The filter configuration description could mention which DR1 fields actually used the optional medium-band filters (e.g., F140M, F182M, F410M, F430M, F480M), since Table 1 does not list medium-band depths and the photometric-redshift quality in fields such as 1420+5252 depends on them.","section":"Sec. 3.3"}],"recommendation":"major_revision","confidential_remarks":"This is a promising survey paper with a valuable public dataset. My major concerns are about the completeness correction underlying the z>13 null result and about the internal consistency of Eq. (1); both are addressable with additional computations or clarifications. The use of a same-author size relation is not a circularity problem in my view, but the lack of independent calibration at z>13 should be explicitly acknowledged and tested. The paper fits the scope of the journal and I do not see grounds for rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is a genuinely useful survey paper. BEACON is a 600-hour NIRCam pure-parallel program, and the DR1 release of 19 fields with 129 z>7 candidates, 11 z>10, and zero z>13 is new data, not a reanalysis. The photometric selection is careful: fixed-aperture S/N cuts, EAzY phot-z with an 80% probability threshold, non-detection checks with two apertures, and a validation against 12 of 13 spectroscopically confirmed z>10 sources. That last point matters; the selection is demonstrably not missing bright high-z galaxies. The number densities at 7<z<13 are consistent with the literature, which is a good sanity check. The paper also ships catalogs and images, which is exactly what a survey paper should do.\n\nThe soft spots are where the reader and the stress-test point. The effective volumes, and hence the z>13 upper limits, come from GLACiAR simulations that assume n=1 Sersic profiles, sizes from a Morishita et al. (2024) relation, and JAGUAR SEDs. At z>12 the SED pool is all JAGUAR spectra combined into one bin, which is a blunt instrument in exactly the redshift range where the null result is claimed. If real z>13 galaxies have higher LyC escape or are more compact, recovery fractions would drop and the upper limits would move up. That is a real caveat, but it is not a fatal flaw: the paper states it openly, the upper limits are already consistent with most literature (the only tension is Robertson et al.'s bright z=14 source), and the authors explicitly note their z>10 selection may be too conservative. In other words, the caveat is proportionate to the claim.\n\nThe only thing I would push back on is over-reading the zero-z>13 result as strong evidence for cosmic variance. The volume is large (~1e5 Mpc3) but the completeness in that redshift window is the least well-calibrated part of the paper. The full survey will be more telling. For now this is a solid DR1 paper with an honest null result and a clear path to more.\n\nWho is it for? Anyone working on high-z luminosity functions, cosmic variance, or survey design. It deserves a serious referee; the data release and the survey legacy value alone justify peer review. I would cite it for the number densities and the survey area.","headline":"A well-executed survey paper: new pure-parallel fields and a carefully validated selection, with a z>13 null result that is honestly caveated but rests on simulation inputs that deserve scrutiny.","tokens_in":36619,"tokens_out":2368,"would_cite":true,"duration_ms":24026,"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":"No galaxy candidates at z>13 found across 19 independent JWST fields","keywords":["BEACON survey","pure-parallel imaging","Lyman-break galaxies","high-redshift galaxies","cosmic variance","JWST NIRCam","galaxy number densities","reionization"],"falsifier":"Recompute the $z>13$ effective volume using simulated galaxies that are a factor of two smaller and redder than the default profiles and templates; if that volume drops substantially, the zero-candidate result no longer constrains cosmic variance.","tokens_in":35226,"feed_emoji":"🔭","tokens_out":15339,"duration_ms":130665,"temperature":0.7,"pith_summary":"BEACON is a JWST Cycle 2 program that uses pure-parallel NIRCam imaging — taking data on a random nearby field while another instrument observes the primary target — to survey roughly a hundred independent sightlines with minimal cosmic variance. In the first 19 fields, covering about $180\\,\\mathrm{arcmin}^2$ and a comoving volume near $10^5\\,\\mathrm{Mpc}^3$, the survey photometrically identifies 129 galaxy candidates at $z>7$, including 11 at $z>10$. The number densities inferred at $7<z<13$ are consistent with earlier surveys and with a constant star-formation efficiency model. The central result is the null: no $z>13$ galaxy candidates are found in this large, statistically independent volume. The paper argues this may indicate that the bright $z>13$ sources reported in small legacy fields are enhanced by cosmic variance rather than typical of the early Universe.","feed_headline":"No galaxies seen at z>13 across 19 independent JWST fields","feed_subtitle":"Null result suggests bright early galaxies in small legacy fields may be cosmic-variance outliers.","key_machinery":"The argument rests on three pieces. The first is the pure-parallel observing strategy itself: NIRCam images are taken while another JWST instrument observes the primary target, so the parallel fields cannot be chosen in advance and are statistically independent sightlines. The second is the Lyman-break dropout method, in which a galaxy is selected by requiring a detection redward of the Lyman break and a non-detection in the bluer filters, with three dropout classes (F090W, F115W, F150W) covering $z\\sim7.3$--$9.7$, $z\\sim9.7$--$13$, and $z\\sim13$--$18$. The third is the per-field effective-volume calculation, $V_{\\rm eff} = \\int (dV/dz)\\,P(M_{\\rm recov}, z)\\,dz$, populated by an adaptation of the GLACiAR2 completeness simulation that injects 1200 galaxies per magnitude--redshift bin, with Sersic $n=1$ profiles, sizes drawn from the $M_{\\rm UV}$--size relation, and JAGUAR spectral templates, and then reruns the full detection, photometry, and selection on every field. This recovery-probability machinery converts raw counts into number densities and upper limits, and it is what makes the $z>13$ null statistically meaningful.","core_discovery":"Using the first 19 fields of the BEACON pure-parallel program, the paper establishes a census of ultraviolet-selected galaxies at $z>7$ that is nearly free of cosmic variance. After Lyman-break dropout selection and photometric-redshift filtering with six to eight NIRCam filters, the team catalogues 129 candidates: 118 F090W dropouts spanning roughly $z\\simeq7.3$--$9.7$, 11 F115W dropouts at $z\\simeq9.7$--$13$, and zero F150W dropouts at $z\\simeq13$--$18$. The sample includes 11 galaxies at $z>10$ and several UV-luminous sources with $M_{\\rm UV}<-21$ mag at $z\\sim8$. The number densities at $7<z<13$ are consistent with previous measurements and with the constant star-formation efficiency model. The paper's central result is that, despite an effective volume of roughly $10^5\\,\\mathrm{Mpc}^3$, no $z>13$ candidate survives, and the resulting upper limits are consistent with previous surveys except at the bright end of one previous measurement containing the source GS-z14-0; the paper interprets this as evidence that the bright $z>13$ sources in legacy fields may be affected by cosmic variance.","pith_inferences":["A testable consequence left implicit by the paper: if the $z>13$ null is real, the UV luminosity function must turn over faster than the extrapolation of the $z\\sim11$ Schechter function, and star-formation models would need to reduce the predicted abundance of bright galaxies.","BEACON's DR1 fields could be combined with legacy fields in a joint likelihood that treats cosmic variance as a free parameter, directly estimating how much of the early JWST tension over bright galaxies is field sampling.","Spectroscopic confirmation of the DR1 $z\\sim8$ UV-luminous candidates would check whether the bright end at $z\\sim8$ is genuinely as populated as the number densities imply, or whether it is contaminated by lower-redshift line emitters.","Applying the same effective-volume machinery to previously published $z>13$ candidates in legacy fields would test whether the discrepancy with the BEACON null survives completeness corrections."],"forward_implications":["If the null persists across the full survey, the bright end of the $z>13$ luminosity function is lower than the abundance of sources like GS-z14-0 in legacy fields would suggest.","The full BEACON sample of roughly 100 fields will reduce cosmic variance in the bright-end UV luminosity function at $z\\sim12$ to about 5 percent, turning field-to-field scatter into a measured quantity rather than a dominant uncertainty.","The 129 DR1 candidates, including 11 at $z>10$, provide a target list for JWST spectroscopy that can confirm redshifts and measure the stellar populations of some of the first galaxies.","The survey design shows that pure-parallel observations can deliver legacy-quality catalogs for lower-redshift science, including massive quiescent galaxies at $z\\sim2$ and brown dwarfs in the Milky Way."],"supporting_citations":[{"why":"Supplies the GLACiAR2 completeness simulation code and the effective-volume method used to convert counts into number densities.","marker":"Leethochawalit et al. (2023)"},{"why":"Provides the M_UV-size relation used to assign sizes to the simulated injected galaxies in the completeness simulation.","marker":"Morishita et al. (2024b)"},{"why":"Provides the JAGUAR spectral templates from which the injected galaxies' spectral energy distributions are drawn.","marker":"Williams et al. (2018)"},{"why":"Provides the cosmic variance calculator used to design the survey and to quantify that BEACON reduces cosmic variance to about 5 percent at z~12.","marker":"Trapp & Furlanetto (2020)"},{"why":"Reports a five-field survey finding no luminous z>10 galaxies, the closest comparison for BEACON's claim that bright high-z sources vary between sightlines.","marker":"Willott et al. (2024)"},{"why":"Contains the luminous z=14 galaxy GS-z14-0, the bright-end measurement with which BEACON's z>13 upper limits are inconsistent.","marker":"Robertson et al. (2024)"},{"why":"Theoretical luminosity function model used to predict expected source counts and to compare the measured number densities at 7<z<13.","marker":"Mason et al. (2023)"},{"why":"Data reduction and photometric redshift approach on which the BEACON pipeline and dropout selection are based.","marker":"Morishita & Stiavelli (2023)"},{"why":"The EAzY code used to fit photometric redshifts and to compute the redshift probability distributions that filter the dropout candidates.","marker":"Brammer et al. (2008)"}],"fun_headline_variants":["JWST finds zero galaxies at z>13 across 19 independent fields","No z>13 galaxies in 19 JWST fields: cosmic variance?","BEACON survey: 19 fields, zero z>13 candidates","Pure-parallel JWST: no z>13 galaxies in huge volume"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The number densities and the $z>13$ upper limits assume that the simulated galaxies used to calibrate each field look like real $z>7$ galaxies in size and color; if real galaxies are much more compact or much redder than the assumed shapes, the recovery fractions and volumes could shift enough to change the null.","fun_headline_variants_meta":{"raw":{"variants":["JWST finds zero galaxies at z>13 across 19 independent fields","No z>13 galaxies in 19 JWST fields: cosmic variance?","BEACON survey: 19 fields, zero z>13 candidates","Pure-parallel JWST: no z>13 galaxies in huge volume"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000768,"raw_usage":{"total_tokens":3523,"prompt_tokens":1184,"completion_tokens":2339,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":800,"completion_tokens_details":{"reasoning_tokens":2269}},"tokens_in":800,"tokens_out":2339,"duration_ms":15256,"temperature":1.0,"reasoning_tokens":2269,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:38:13.748483+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the $z>13$ effective volume using simulated galaxies that are a factor of two smaller and redder than the default profiles and templates; if that volume drops substantially, the zero-candidate result no longer constrains cosmic variance.","supporting_citations":[],"review_version":1}