{"id":"4ad2535b-38a7-4984-aef7-7627a07262e6","arxiv_id":"1908.04310","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"With updated simulations and a Bayesian selection method, the authors forecast that Euclid will discover over 100 quasars at redshift 7 to 7.5 and about 8 at redshift greater than 8, depending on how fast the quasar population declines with redshift.","lead":"This paper predicts how many very distant quasars, the bright beacons powered by supermassive black holes in the early universe, the Euclid space telescope should find in its wide sky survey. Using improved simulations, it estimates over 100 quasars at redshift 7 to 7.5 and about 8 beyond redshift 8 if the quasar population declines steeply with cosmic time.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The k=-0.92 yield in Table 3 is computed with BMC selection functions whose quasar prior is fixed to k=-0.72 (Sect. 3.2.1); the resulting overestimate of the selection function is unquantified and could affect the headline >100 count.","rationale":"The paper is a transparent forecasting study that correctly identifies the QLF decline rate k as the dominant physical uncertainty and explicitly labels k=-0.92 as a pessimistic scenario, so the single-power-law extrapolation is an acknowledged limitation rather than a hidden flaw. The internal prior mismatch, however, is not acknowledged and directly affects the numerical headline: the k=-0.92 yield is produced by integrating a k=-0.92 QLF over selection functions that were computed assuming k=-0.72. Because the BMC prior enters Pq directly, this mismatch biases the selection function toward higher completeness precisely in the faint J~23 regime where many of the predicted quasars lie. This is a concrete, testable technical issue, and it is more load-bearing than the acknowledged external uncertainties such as the eventual LSST/Pan-STARRS depths or quasar template diversity, which the paper already discusses or tests in Sect. 5.6. The reader's conditional verdict already calls for recomputing or justifying the BMC selection functions for the k=-0.92 prior; this stress test agrees and sharpens the reason: without that check, the k=-0.92 columns in Table 3 do not establish the headline yield under self-consistent assumptions. A corrected calculation may still yield more than 100 quasars, but the current table does not demonstrate it, so the verdict remains CONDITIONAL pending that check.","tokens_in":36075,"tokens_out":9327,"duration_ms":98673,"concrete_test":"Recompute the BMC selection function for the ground-based optical scenario with the quasar surface-density prior in Eq. (3) set to the same Jiang et al. (2016) QLF but with k=-0.92, leaving all contaminant models and the Pq=0.1 threshold unchanged; then re-integrate the k=-0.92 QLF over this new selection function. Compare the resulting 7.0<z<7.5 count with the Table 3 entry of 117. If the count drops below 100 (or by more than about 10%), the headline yield must be revised or explicitly stated to depend on a k=-0.72 prior. As a control, repeat the same procedure with k=-0.72 to verify that the pipeline reproduces the existing table.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the Bayesian model comparison (Sect. 3.1.1), the quasar weight Wq is proportional to the quasar surface-density prior, which Sect. 3.2.1 fixes to the Jiang et al. (2016) QLF extrapolated with k=-0.72. The selection functions (Fig. 8) are therefore computed for a k=-0.72 world. Section 4.1 then integrates a k=-0.92 QLF over these same selection functions to produce the k=-0.92 columns of Table 3. If the true evolution were k=-0.92, the prior used in the selection step should also be k=-0.92; the adopted prior is too high by factors of about 1.8 at z=7.25 and 2.9 at z=8.25 (10^{0.2(z-6)}). A higher quasar prior inflates Pq for sources that resemble quasars, so the selection function is more complete than the self-consistent k=-0.92 selection function. The effect is largest near J~23, where quasars and contaminants are both selected near the Pq=0.1 threshold and where a substantial fraction of the Table 3 counts reside. Thus the k=-0.92 columns are not self-consistent forecasts, and the headline statement that Euclid 'should nevertheless find over 100 quasars' at 7.0<z<7.5 could be an overestimate. The paper does not quantify this bias; it is distinct from the acknowledged uncertainty in the true value of k.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents simulated predictions of the number of 7<z<9 quasars that the Euclid wide survey will discover. The authors update the Red Book calculation using revised NISP filter curves, the Jiang et al. (2016) z=6 quasar luminosity function extrapolated with k=-0.72 or k=-0.92, improved models of MLT dwarf and early-type galaxy contaminants, and a Bayesian model comparison (BMC) selection method alongside a minimum-chi-squared method. They compute selection functions and integrate them over the QLF to obtain yields in redshift bins. With ground-based z-band data, they predict over 100 quasars at 7.0<z<7.5 even for k=-0.92, about 25 at z>7.5, and about 8 at z>8.0. They also discuss contamination rates, follow-up feasibility, and the timeline of Euclid data releases, and estimate that k could be measured to 1-sigma uncertainty of 0.07 over 7<z<8.","tokens_in":36418,"tokens_out":7814,"duration_ms":75933,"significance":"If the predictions are correct, Euclid will provide the first large statistical sample of z>7 quasars, enabling direct measurement of the quasar luminosity function at 7<z<9, constraints on SMBH growth, and Ly-alpha damping-wing measurements of reionization. The paper is thorough and reproducible in structure: the population models are described in detail, the BMC and chi-squared methods are compared, and the sensitivity to contaminant populations (Sects. 5.4-5.5) and quasar template variations (Sect. 5.6) is explicitly tested. The predicted yield for the k=-0.92 scenario, however, rests on an internal inconsistency in the selection-function prior, which needs to be resolved before the headline numbers can be considered robust.","major_comments":[{"comment":"The k=-0.92 yield forecasts are computed by integrating the k=-0.92 QLF over selection functions that were themselves derived with a quasar prior fixed to k=-0.72 (stated in Section 3.2.1). Because P_q in Eq. (2) depends on the quasar surface-density prior in Eq. (3), this is not a self-consistent forecast: in a true k=-0.92 world the prior would be lower by a factor of 10^{0.2(z-6)} (about 1.8 at z=7.25 and 2.9 at z=8.25), which would lower P_q for marginal sources and likely reduce the faint-end completeness. Since a substantial fraction of the Table 3 counts come from near J~23, the headline claim of 'over 100 quasars' for k=-0.92 in the abstract and Section 4.1 could be an overestimate. Please recompute the selection functions for the k=-0.92 prior or quantify the resulting bias; the current treatment leaves the size of the effect unquantified.","section":"Sections 3.2.1 and 4.1, Table 3"}],"minor_comments":[{"comment":"The yields are reported as point estimates with no uncertainties; adding at least Poisson errors (and a brief discussion of systematic uncertainties from the contaminant surface densities) would help the reader assess the significance of the differences between scenarios.","section":"Table 3"},{"comment":"The statement that 'the first Euclid quasars at z>7.5 should be found in DR1' is stronger than the numbers justify; for k=-0.92 the predicted DR1 yield at z>7.5 is 2.3 sources, so 'expected' or 'could' would be more appropriate.","section":"Section 5.1, Table 4"},{"comment":"The headline 'over 100 quasars with 7.0<z<7.5' for k=-0.92 refers to the case with ground-based z-band data; the abstract mentions that z-band data improve selection over 7<z<8, but the claim should state this condition explicitly at that point.","section":"Abstract"},{"comment":"The notation 'theta_t is the set of parameters describing a single population' is ambiguous because the quasar population is described by a continuous grid in M_1450 and z while the MLT population is a set of discrete spectral types; please clarify.","section":"Section 3.1.1, Eq. (3)"},{"comment":"The statement that the results are insensitive to the near-zone size is not demonstrated; a one-sentence sensitivity test or a reference would support this claim.","section":"Section 3.2.1"}],"recommendation":"major_revision","confidential_remarks":"This Euclid Collaboration paper is likely to attract wide attention because it quantifies the mission's potential for z>7 quasar science. The main technical issue is the internal inconsistency in the k=-0.92 selection-function prior, which directly affects the headline yield and should be fixed or explicitly bounded before acceptance. The paper is otherwise careful and includes useful cross-checks (chi-squared method, contaminant simulations, template variation tests). I see no need for additional external validation beyond what is already presented, but the authors should address the prior mismatch in the revised version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe short version: this is a careful, useful update of the Euclid Red Book forecast for z>7 quasars, with genuinely better selection modeling and useful yield tables. But the k=-0.92 scenario has an internal inconsistency that the authors should fix before I'd trust the headline 'over 100' claim.\n\nWhat's new and good: the paper replaces heuristic colour cuts with a Bayesian model comparison (BMC) method, updates the NISP filter wavelengths, uses better MLT dwarf and compact early-type galaxy contamination models, and extends the forecast down to J~23 and over 7<z<9. The selection functions are clearly described, the authors test several modeling choices (contaminant populations, early-type assumptions, quasar SED variations), and the conclusion that k, the QLF decline rate, is the dominant uncertainty is honest and correct. The tables and figures will be a useful reference for the community planning Euclid follow-up.\n\nThe soft spot: the BMC selection functions are computed with a quasar prior fixed to k=-0.72 (Sect. 3.2.1), but then Table 3 and the abstract use those same selection functions to integrate a k=-0.92 QLF. If the true decline were steeper, the prior should be lower; the k=-0.72 prior is higher by a factor ~1.8 at z=7.25 and ~2.9 at z=8.25. That inflates Pq for marginal sources near J~23 and makes the k=-0.92 selection function more complete than a self-consistent one would be. The paper never quantifies this, so the k=-0.92 yields, including the headline 'over 100' at 7.0<z<7.5, are likely overestimated by some unknown amount, my guess tens of percent. The chi-squared method, which uses no prior, gives similar numbers at z>8, so the issue mainly affects 7<z<8.\n\nSmaller issues: Table 3 has no propagated uncertainties; for a forecast paper that is a minor annoyance, not a fatal flaw. The ground-based z-band scenario depends on assumed Pan-STARRS/LSST depths that are not yet secured; the paper does label these as working assumptions.\n\nBottom line: the paper deserves a serious referee. I would recommend acceptance after the authors either recompute the selection functions for k=-0.92 or demonstrate that the prior mismatch has a negligible effect on the yields. Also ask them to add error bars or state clearly that the numbers are central values only.\n\nWorth citing if you work on high-z quasars or Euclid survey forecasts.","headline":"A careful, transparent update of the Euclid z>7 quasar yield forecast, but the k=-0.92 scenario has an unquantified internal inconsistency (selection functions built with a k=-0.72 prior) that likely makes the headline 'over 100' an overestimate.","tokens_in":37726,"tokens_out":6001,"would_cite":true,"duration_ms":60211,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Euclid's wide survey should yield over 100 quasars at redshifts 7.0-7.5, and roughly 25 beyond 7.5, even if quasar numbers fade fast.","keywords":["high-redshift quasars","quasar luminosity function","Euclid wide survey","Bayesian model comparison","reionization","Lyman-alpha damping wing","near-infrared selection","MLT dwarfs"],"falsifier":"Search for spectroscopically confirmed quasars at $7.5<z<8.0$ in the Euclid DR1 southern field (1250 deg2 with one-year LSST data), where the prediction is 4.1 quasars for $k=-0.72$ and 1.8 for $k=-0.92$; a confirmed count consistent with zero across the full DR1 area would falsify the yield forecast.","tokens_in":35867,"feed_emoji":"🔭","tokens_out":12760,"duration_ms":112576,"temperature":0.7,"pith_summary":"This paper predicts how many quasars at redshifts $7<z<9$ the Euclid wide survey should find, updating an earlier estimate with revised near-infrared filters, steeper assumed declines of the quasar luminosity function, better models of contaminating stars and galaxies, and a Bayesian selection method. The headline forecast is that even in the pessimistic case where the decline of quasar number density accelerates beyond $z=6$ ($k=-0.92$), Euclid should discover over 100 quasars at $7.0<z<7.5$ and about 25 beyond the current record redshift $z=7.5$, including roughly 8 at $z>8$. Adding deep $z$-band data from ground-based surveys roughly doubles the yield over $7<z<8$, and the first $z>7.5$ quasars are predicted to appear in the DR1 data release expected in 2024. If these numbers hold, Euclid will provide the first large statistical sample of $z>7$ quasars, enabling direct measurements of the quasar luminosity function and new probes of cosmic reionisation through Lyman-$\\alpha$ damping wings.","feed_headline":"Euclid forecast: 100 quasars at redshift 7, 25 beyond 7.5","feed_subtitle":"Even if quasar numbers fade fast, Euclid should break the redshift-7.5 record and probe reionization.","key_machinery":"The machinery is the Bayesian model-comparison (BMC) selection function. For simulated quasars on a grid of luminosity and redshift, the method computes a posterior quasar probability $P_q$ from Gaussian photometric likelihoods weighted by the surface densities of three populations: quasars from the $z=6$ luminosity function, MLT dwarfs from published luminosity functions and colours, and compact early-type galaxies at $z=1-2$ modelled from COSMOS data with a size-mass relation. The selection function records the fraction of quasars with $P_q>0.1$ as a function of absolute magnitude and redshift, and the predicted yields follow by integrating the assumed luminosity function over these functions. The key comparison that carries the argument is between using Euclid's broad $O$ band and using deep ground-based $z$-band data, which sharpens the contrast across the Lyman break and deepens the selection by about a magnitude over $7<z<8$.","core_discovery":"The central result is a new set of Euclid quasar selection functions, derived by simulating quasars on a grid of absolute magnitude and redshift and recording the fraction that survive a Bayesian model-comparison selection with quasar probability threshold $P_q>0.1$. When the $z=6$ quasar luminosity function is extrapolated as $\\Phi\\propto 10^{k(z-6)}$, the predicted yields are 204 quasars at $7.0<z<7.5$ and 45 at $7.5<z<8.0$ for $k=-0.72$ with ground-based optical data, falling to 117 and 19 for $k=-0.92$; the corresponding numbers beyond $z=8$ are 23 and 8. With deep $z$-band data the selection reaches $J_{AB}\\sim23$, about a magnitude deeper than with Euclid's own optical band over $7<z<8$, and at that limit the predicted contamination leaves a selection efficiency of roughly two-thirds. The paper further shows that quasars at $z>8$ can be selected from Euclid $OYJH$ photometry alone, and that $k$ can be recovered to a $1\\sigma$ uncertainty of 0.07 over $7<z<8$ if $k=-0.72$.","pith_inferences":["Inference: if the predicted counts materialise, the same sample could be used to test black-hole seeding scenarios by checking whether the supply of $z>7$ bright quasars is consistent with Eddington-limited growth from stellar-mass seeds, because the survey would measure the bright-end space density directly rather than extrapolating it.","Inference: the forecast's sensitivity to $z$-band depth implies that the realised yield is partly a survey-coordination outcome: a one-magnitude loss in Pan-STARRS or LSST coverage would more than halve the $7<z<8$ yield, so the paper functions as a quantitative argument for prioritising the ground-based overlap.","Inference: because the simulation excludes gravitationally lensed quasars with optical flux from the deflecting galaxy, a future search that adds lensed templates could find an additional population; counting lensed candidates in the final sample would test whether the empirical lensing fraction near 1% holds at $z>7$.","Inference: a quick test of the selection before DR1 is possible with the Q1 release over about 50 deg2, where the model predicts at most one $7<z<9$ quasar but also predicts the expected number of false candidates; matching that contamination rate would calibrate the Bayesian priors for later releases."],"forward_implications":["Under the nominal $k=-0.72$ decline, Euclid should yield over 200 quasars at $7.0<z<7.5$ and 45 at $7.5<z<8.0$ with ground-based optical data, giving the first large sample of $z>7$ quasars.","Even under the steeper $k=-0.92$ decline, more than 100 quasars at $7<z<7.5$ and about 8 beyond $z>8$ are expected, so Euclid should break the current redshift record.","Euclid's samples will constrain the bright-end slope of the quasar luminosity function over $7<z<8$ with 8m telescopes, while JWST or E-ELT follow-up will be needed to measure the faint-end slope.","The first quasars at $z>7.5$ should appear in the DR1 data release in 2024, with more than ten $7<z<9$ quasars predicted over the 1250 deg2 southern DR1 area with LSST data.","Assuming $k=-0.72$, the decline parameter $k$ can be measured to a $1\\sigma$ uncertainty of 0.07 over $7<z<8$."],"supporting_citations":[{"why":"Supplies the $z=6$ quasar luminosity function and the $k=-0.72$ decline rate that anchor all predicted yields.","marker":"Jiang et al. 2016"},{"why":"Defines the Bayesian model-comparison method, including the $P_q=0.1$ selection threshold.","marker":"Mortlock et al. 2012"},{"why":"The Red Book calculation whose yields, depths, and filter assumptions this paper updates.","marker":"Laureijs et al. 2011"},{"why":"Provides L- and T-dwarf number densities and colours used to model the dominant contaminating population.","marker":"Skrzypek et al. 2016"},{"why":"Supplies the M-dwarf luminosity function used to set M0-M6 number densities.","marker":"Bochanski et al. 2010"},{"why":"The COSMOS catalogue from which the early-type galaxy surface density is fitted and the contaminant model is validated.","marker":"Laigle et al. 2016"},{"why":"Gives the size-mass relation used to treat faint early-type galaxies as compact point sources.","marker":"van der Wel et al. 2014"},{"why":"Supplies the minimum-$\\chi^2$ SED-fitting method and cuts that are compared against the Bayesian approach.","marker":"Reed et al. 2017"},{"why":"Provides the stellar population synthesis models used to compute early-type galaxy colours for two formation redshifts.","marker":"Bruzual & Charlot 2003"},{"why":"Defines the Pan-STARRS survey whose $z$-band depth sets one of the two ground-based optical scenarios.","marker":"Chambers et al. 2016"}],"fun_headline_variants":["Euclid to break quasar redshift record with 25 new finds","Quasar hunt: Euclid to find 100+ at z>7 and 8 beyond z>8","Euclid's quasar forecast: 100+ at z~7, 8 at z>8","Euclid to probe reionization era with 100+ new quasars"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The forecast rests on the assumption that the measured decline of bright quasars between redshifts 5 and 6 continues unchanged all the way to redshift 9 as a single power law, $\\Phi\\propto 10^{k(z-6)}$, with the decline rate set to $k=-0.72$ or $k=-0.92$; if the number density falls faster or the luminosity function bends, every predicted count shifts.","fun_headline_variants_meta":{"raw":{"variants":["Euclid to break quasar redshift record with 25 new finds","Quasar hunt: Euclid to find 100+ at z>7 and 8 beyond z>8","Euclid's quasar forecast: 100+ at z~7, 8 at z>8","Euclid to probe reionization era with 100+ new quasars"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001435,"raw_usage":{"total_tokens":5970,"prompt_tokens":1315,"completion_tokens":4655,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":931,"completion_tokens_details":{"reasoning_tokens":4558}},"tokens_in":931,"tokens_out":4655,"duration_ms":29143,"temperature":1.0,"reasoning_tokens":4558,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:46:10.021335+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search for spectroscopically confirmed quasars at $7.5<z<8.0$ in the Euclid DR1 southern field (1250 deg2 with one-year LSST data), where the prediction is 4.1 quasars for $k=-0.72$ and 1.8 for $k=-0.92$; a confirmed count consistent with zero across the full DR1 area would falsify the yield forecast.","supporting_citations":[{"cited_title":"D., Fan, X., et al","cited_arxiv_id":null,"evidence_quote":"Supplies the $z=6$ quasar luminosity function and the $k=-0.72$ decline rate that anchor all predicted yields."},{"cited_title":"J., & Faherty, J","cited_arxiv_id":null,"evidence_quote":"Provides L- and T-dwarf number densities and colours used to model the dominant contaminating population."},{"cited_title":"J., Hawley, S","cited_arxiv_id":null,"evidence_quote":"Supplies the M-dwarf luminosity function used to set M0-M6 number densities."}],"review_version":1}