{"id":"ed21a2b7-0d1c-4c3c-98da-e1befa258255","arxiv_id":"1908.00731","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Galaxies produce a 3.4 micron cosmic infrared background of 9.0 +/- 0.5 kJy/sr, derived by integrating a new luminosity function and mean galaxy SED.","lead":"Using a new census of galaxy brightness and an average galaxy spectrum, this paper estimates that galaxies contribute 9.0 plus or minus 0.5 kJy per steradian to the 3.4 micron infrared sky background. The value is higher than many earlier galaxy-only estimates and sits between direct measurements of the total background and gamma-ray upper limits.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 9.0 kJy/sr headline is set by a single preferred MCMC chain; the comparably motivated High z Trim Prior chain gives ~5 kJy/sr, so the quoted statistical (0.5) and systematic (2 kJy/sr) bounds are not supported by the internal model spread.","rationale":"The reader's weakest-assumption analysis identifies the same fragile point: the EBL integral inherits the Schechter LF evolution model (Eqs. 8-10) and, in practice, one preferred posterior chain, while a second preferred chain with the same data and parameterization gives a factor-of-two lower background. My reading of the full text reinforces this: (i) the exclusion of High z Trim Prior is argued from physical plausibility rather than a formal model comparison; (ii) about half of the reported 9 kJy/sr comes from z>1, beyond the redshift range of the LF measurement, where the constant-alpha and non-evolving-SED assumptions are acknowledged limitations; and (iii) the number-count validation is not decisive because the paper itself flags the High z Prior model's faint-end W1 over-prediction as suspect. These points do not invalidate the measurement, but they do mean the quoted 0.5 kJy/sr statistical error and the 'unlikely to be greater than 2 kJy/sr' systematic bound are too narrow relative to the internal chain spread. The data products are public and the chains are published, so the disagreement can be resolved by the concrete reanalysis above. Since the reader already recommended CONDITIONAL acceptance on essentially this basis, my read is a confirmation rather than a change of verdict.","tokens_in":15512,"tokens_out":7714,"duration_ms":78118,"concrete_test":"Using the published MCMC chain files (Figshare doi:10.6084/m9.figshare.4245443), recompute the z<5 EBL integral for the High z Prior chain after applying the High z Trim completeness selection while keeping the identical alpha prior. If the value moves from ~9 to ~5 kJy/sr, the untrimmed low-completeness sample is what sets the headline and the systematic budget must be expanded to at least the between-chain spread; if it remains near 9 kJy/sr, the decisive difference is the interaction of the alpha prior with the trim selection, which should then be tested by recomputing the High z Trim Prior chain with the High z Prior alpha prior.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the selection of the High z Prior chain in Section 3 (Eqs. 8-10) as the basis for the EBL integral in Eq. 5. The paper itself presents two preferred combined chains, High z Prior and High z Trim Prior, whose z<=5 backgrounds differ by roughly a factor of two in Figure 3. The rejection of High z Trim Prior rests on two interpretive statements: its alpha=-1.93+/-0.04 is uncomfortably close to the -2 divergence, and its phi* evolution implies a comoving number density that is currently increasing at 1.9+/-0.7 e-folds per Hubble time. These are plausibility arguments, not a quantitative falsification, and the same data set produced both chains. The issue is compounded by the redshift decomposition: the example MCMC rows in Table 1 show I_nu(z<1) ~ 4.6 kJy/sr while I_nu(z<5) ~ 9.3 kJy/sr, so nearly half of the headline value is an extrapolation beyond z=1, the redshift range of the LF data. The number-count checks in Figures 8-9 are presented as confirmation, but Section 5 states that the High z Prior model over-predicts the faint W1 counts and 'should be viewed as suspect for the purposes of predicting faint galaxy flux counts'; thus the independent check that would validate the faint/high-z extrapolation is weakened exactly where the EBL integral is sensitive. Because the central claim's uncertainty budget ignores the between-chain spread of ~4 kJy/sr, the correctness of the headline depends on a model-selection judgment that the paper does not defend quantitatively.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper estimates the contribution of galaxies to the 3.4 μm extragalactic background light (EBL) by integrating a luminosity function (from Lake et al. 2018) with a mean galaxy SED (from Lake & Wright 2016). The headline result is Iν = 9.0±0.5 kJy/sr for z<5, with a claimed systematic uncertainty below 2 kJy/sr. The authors propagate the full MCMC posterior of the LF parameters through the integral, provide machine-readable posterior chains, and compare predicted source counts in WISE and SDSS bands with observed counts. The paper concludes that known galaxies produce most of the 3.4 μm EBL, consistent with direct total-EBL measurements and blazar upper limits.","tokens_in":15927,"tokens_out":2835,"duration_ms":29810,"significance":"If the headline result holds, this is an important measurement: it is higher than most previous galaxy-integration estimates and helps to close the gap between direct EBL measurements and galaxy counts. The methodology has clear strengths: the EBL integral is a forward calculation from independently measured LF and SED quantities, the MCMC uncertainty propagation is statistically sound, the posterior chains are published under a DOI, and the comparison with source counts at multiple wavelengths is a useful diagnostic. The paper also is unusually honest about model limitations, including the non-evolving SED and constant faint-end slope. These strengths make the central claim worth serious consideration, provided the model-selection issue identified below is resolved.","major_comments":[{"comment":"The choice of the High z Prior chain over the High z Trim Prior chain is the load-bearing step in determining the headline value, but the paper does not provide a quantitative justification. Both chains are explicitly described as preferred, yet their z<5 backgrounds differ by roughly a factor of two (about 9 kJy/sr versus about 5 kJy/sr in Figure 3). The stated reasons for rejecting High z Trim Prior are plausibility arguments: its α = -1.93 is uncomfortably close to the α = -2 divergence, and its implied comoving number density evolution of 1.9±0.7 e-folds per Hubble time is considered too fast. These are not model-independent falsifications, and the same data set produced both chains. As written, the quoted statistical uncertainty of ±0.5 kJy/sr and the systematic bound of 2 kJy/sr do not include this between-chain spread of roughly 4 kJy/sr. The authors should either defend the selection with a quantitative criterion (e.g., a goodness-of-fit comparison or an explicitly motivated prior) or incorporate the chain-to-chain variation into the systematic uncertainty budget.","section":"Section 3; Figure 3"},{"comment":"The redshift decomposition of the result exposes a second load-bearing concern. Table 1 shows Inu(z<1) ≈ 4.6 kJy/sr while Inu(z<5) ≈ 9.3 kJy/sr for the High z Prior chain, meaning that roughly half of the headline signal comes from z>1, beyond the redshift range of the LF data (which are limited to z≲1 as stated in Section 3). The z>1 contribution therefore rests entirely on the adopted extrapolation of φ*(z) and L*(z) in Equations 8-10. Because the competing High z Trim Prior chain effectively changes this extrapolation by a factor of two, the paper should quantify how much of the 9 versus 5 kJy/sr discrepancy arises specifically from the high-redshift extrapolation and should present a sensitivity analysis of the EBL integral to the assumed evolution model. Without such an analysis, the claim 'Iν = 9.0±0.5 kJy/sr' is not supported by the internal model spread.","section":"Section 4; Table 1; Figure 3"},{"comment":"The paper uses the source-count comparisons as confirmation of the model, but the text itself weakens that confirmation precisely where the EBL integral is most sensitive. In Section 5, at the end of the flux-count comparison, the authors state that the High z Prior model's over-prediction at the faint end of the W1 counts 'should be viewed as suspect for the purposes of predicting faint galaxy flux counts.' Since the EBL integral is dominated by faint galaxies and high redshifts, this admission directly undermines the use of Figures 8 and 9 as independent checks of the background estimate. The authors need to explain how the faint-end count excess affects the credibility of the EBL integral, or to perform a quantitative test that separates the flux-count information from the EBL normalization. As it stands, the confirmation argument is circular in a way that is acknowledged but not addressed.","section":"Section 5; Figures 8-9"},{"comment":"The systematic uncertainty budget is asserted rather than derived. The sentence 'any modification from the true value caused by the systematic limitations in this work is unlikely be more than about 2 kJy/sr in either direction' is not supported by a quantitative error analysis. The acknowledged limitations include a non-evolving mean SED, a constant faint-end slope, and an uncertain high-redshift L*(z) evolution, each of which could shift the integral by amounts comparable to or larger than the quoted 2 kJy/sr. The authors should either provide a systematic error estimate that combines these effects (e.g., by reweighting the chains or by marginalizing over plausible model extensions) or replace the 2 kJy/sr claim with a more defensible bound.","section":"Section 5; last paragraph of Discussion"}],"minor_comments":[{"comment":"In the phrase 'α = -1.93±-0.04' the sign of the error is garbled; it should read 'α = -1.93±0.04'.","section":"Section 4, paragraph describing Figure 3"},{"comment":"The notation for LSED and µν is intricate; a short remark explicitly stating that µν is the mean of the normalized SED at fixed Lf and z would help readers who are not already familiar with the spectro-luminosity functional formalism from Lake et al. (2017).","section":"Section 2, Equation (5) and surrounding text"},{"comment":"The description of the 'High z' versus 'High z Trim' samples is given only in the main text; repeating the definitions in the figure caption would make Figure 3 more self-contained.","section":"Section 4, Figure 3 caption and text"},{"comment":"The caption does not state that the literature points are adjusted to 3.4 μm assuming Iν is roughly constant, although this is mentioned in the text. Adding this to the caption would prevent misreading.","section":"Section 5, Figure 7 caption"},{"comment":"The WMAP9 parameter URL is given as a footnote; it would be more robust to cite the Hinshaw et al. (2013) parameters in the text as well, particularly because the cosmology affects Dc(z) in Equation (5).","section":"Section 1, footnote 2"}],"recommendation":"major_revision","confidential_remarks":"The concern raised in the stress-test note is real and lands squarely on the central claim. The paper itself presents two preferred chains whose z<5 backgrounds differ by about a factor of two, and the rejection of one chain is based on plausibility rather than a quantitative test. This is not a reason to reject the work, but it is a reason to require a major revision that either justifies the chain selection or inflates the quoted systematic uncertainty to encompass the model spread. The paper's methodological strengths (forward modeling, MCMC propagation, published chains, multi-wavelength count comparisons) are substantial, and the result is scientifically valuable if the systematic issue is resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper makes a genuinely new measurement: it integrates a luminosity function measured from scratch (using spectroscopic redshifts and WISE photometry) with a mean galaxy SED to get the 3.4 μm contribution to the EBL. The math is standard, the public data products are a real asset, and the forward-integral approach is not fit to the EBL itself, so there is no direct circularity. The authors are also unusually transparent about the limitations of their model — constant SED, constant faint-end slope, and the suspect faint-end number counts. Credit where due.\n\nThe soft spot is exactly where the stress-test note lands. The headline 9.0 kJy/sr comes from one preferred MCMC chain, 'High z Prior'. Another combined chain, 'High z Trim Prior', which the paper itself treats as a serious alternative, gives about 5 kJy/sr. The rejection of the lower chain rests on plausibility arguments — its faint-end slope is close to -2, and its comoving number density evolution looks odd — but those are not quantitative falsifications, and both chains came from the same data. The between-chain spread is ~4 kJy/sr, yet the quoted systematic bound is 2 kJy/sr. That bound is asserted, not derived, and the paper's own discussion contradicts it: the High z Prior model over-predicts faint W1 counts, and the text says that model 'should be viewed as suspect for the purposes of predicting faint galaxy flux counts.' That is exactly the regime where the EBL integral is sensitive, since nearly half the headline value comes from z > 1, beyond the spectroscopic redshift range of the LF data.\n\nSo the central claim is not robust to a reasonable choice of model. This does not make the paper worthless. It is an important data point in the 1–4 μm EBL debate, and the method is clear enough to be reproduced and challenged. But the proper conclusion is something like 'the galaxy contribution is probably between 5 and 9 kJy/sr, with the exact value depending on how you treat the faint end and high-redshift extrapolation,' not a 9.0±0.5 measurement with a 2 kJy/sr systematic cap.\n\nThe paper deserves serious peer review. It is substantively important, the data products are public, and the analysis is auditable. A good referee should demand a defensible model-selection argument or a systematic budget that reflects the real between-chain scatter. I would bring it to a reading group as a case study in how model choice drives a headline result.\n\nRecommendation: send to a competent referee, but expect that referee to push for major revisions on the uncertainty quantification.","headline":"The 9 kJy/sr galaxy EBL at 3.4 microns is a real, reproducible measurement, but the paper's own alternative preferred chain gives ~5 kJy/sr and the quoted 2 kJy/sr systematic bound does not cover that spread, so the headline is less solid than the abstract implies.","tokens_in":16461,"tokens_out":1946,"would_cite":true,"duration_ms":22256,"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":"This paper claims that galaxies contribute $I_\\nu = 9.0\\pm0.5$ kJy/sr to the 3.4 μm extragalactic background light, higher than most prior galaxy-integration estimates but consistent with direct measurements and blazar upper limits.","keywords":["extragalactic background light","cosmic infrared background","luminosity function","spectral energy distribution","WISE","near-infrared","galaxy evolution","galaxy statistics"],"falsifier":"A deep spectroscopic survey measuring the faint-end slope of the 2.4 μm luminosity function at $z \\approx 0.5$–$1$ to an uncertainty of $\\pm0.1$ would discriminate: if the slope is as steep as $\\alpha \\lesssim -1.9$, the predicted background falls to roughly $5\\,\\mathrm{kJy\\,sr^{-1}}$. Alternatively, a fully completeness-corrected W1 galaxy count at the faint end would test the model directly, since the paper's own comparison shows the preferred model over-predicting the faint counts.","tokens_in":15342,"feed_emoji":"🌌","tokens_out":15457,"duration_ms":133785,"temperature":0.7,"pith_summary":"The paper sets out to determine how much of the extragalactic background light (EBL) at 3.4 μm was emitted by galaxies. It combines a recently measured galaxy luminosity function at 2.4 μm with an average galaxy spectrum, then integrates the product over cosmic time. The central result is $I_\\nu = 9.0 \\pm 0.5\\,\\mathrm{kJy\\,sr^{-1}}$ for the galaxy contribution, with systematic errors estimated at no more than about $2\\,\\mathrm{kJy\\,sr^{-1}}$. That value is higher than most previous galaxy-integration estimates, but agrees with direct measurements of the total background and with gamma-ray blazar upper limits. If the result is right, known galaxies produce the bulk of the 3.4 μm background, leaving little room for an unidentified source at this wavelength.","feed_headline":"Galaxies supply most of the 3.4-micron cosmic background light","feed_subtitle":"WISE-based galaxy census puts the galaxy contribution at 9 kJy/sr, matching direct sky measurements and blazar limits.","key_machinery":"The central object is the spectro-luminosity functional, which describes how many galaxies occupy each combination of luminosity and spectral energy distribution. The paper never needs the full functional: it uses its first moment, the comoving spectral luminosity density $\\rho_\\nu$, computed from a Schechter luminosity function $\\Phi(L_f,z)$ (a power-law faint end with an exponential cutoff at a characteristic luminosity) and the mean normalized galaxy SED $\\mu_\\nu$. The load-bearing identity is Equation (5), $d^2 I_\\nu / (dz\\,dL) = [\\mu_\\nu([1+z]\\nu, L_f, z) / (\\Omega_{\\rm sky}(1+z))]\\,(dD_c/dz)\\,L_f\\,\\Phi(L_f,z)$, which converts the luminosity function into a contribution to the background at each redshift and luminosity. Integrating this over $z$ and $L$ gives $I_\\nu$. The luminosity function is an evolving Schechter function whose parameters come from posterior chains of a previous measurement, and the mean SED comes from a separate fit to deep survey photometry; the full posterior distribution of the LF parameters is propagated through the integral to get the uncertainty.","core_discovery":"The paper claims that the contribution of galaxies to the 3.4 μm extragalactic background light is $I_\\nu = 9.0 \\pm 0.5\\,\\mathrm{kJy\\,sr^{-1}}$ ($\\nu I_\\nu = 8.0 \\pm 0.4\\,\\mathrm{nW\\,m^{-2}\\,sr^{-1}}$ per e-fold), a value higher than most earlier galaxy-integration estimates. The estimate is built by combining a 2.4 μm galaxy luminosity function, measured from more than half a million galaxies with spectroscopic redshifts, with a mean galaxy spectral energy distribution derived from deep multi-band photometry, and integrating the product over redshift to $z=5$. The paper treats the luminosity-function parameters as a Bayesian posterior and propagates the full chain through the integral, so the quoted error reflects correlated parameter uncertainties. It argues that systematic uncertainties are unlikely to exceed $2\\,\\mathrm{kJy\\,sr^{-1}}$ in either direction, and that the result is consistent with direct sky measurements of the total background and with upper limits set by TeV blazar observations.","pith_inferences":["Editorial inference: If the 9 kJy/sr galaxy contribution survives deeper tests, the 3.4 μm band becomes a place where direct sky measurements, galaxy integration, and gamma-ray opacity constraints agree, which would localize the unresolved near-infrared background controversy to shorter wavelengths.","Editorial inference: The published evolving EBL tables could be combined with future TeV blazar spectra to test whether the gamma-ray horizon is explained by galaxies alone or requires extra light from, for example, extended galactic halos.","Editorial inference: The faint end is the decisive measurement: if the true faint-end slope $\\alpha$ of the luminosity function is near $-2$, the integral requires a low-luminosity cutoff, and the value of that cutoff—not just the slope—determines how much faint-galaxy light is missed."],"forward_implications":["At 3.4 μm, known galaxies contribute $I_\\nu = 9.0 \\pm 0.5\\,\\mathrm{kJy\\,sr^{-1}}$, so ordinary galaxies—not an unidentified source—account for most of the extragalactic background at this wavelength.","This galaxy-only value is consistent with both direct measurements of the total EBL and upper limits from TeV blazars, easing the earlier tension among these methods near 3.4 μm.","The evolving luminosity density and background tables published with the paper can be used directly to compute the optical depth of the universe to very-high-energy gamma rays.","If the model's over-prediction of faint W1 counts is real, deeper completeness-corrected counts will push the estimate toward the 5 kJy/sr alternative; if incompleteness explains the discrepancy, the 9 kJy/sr value will stand."],"supporting_citations":[{"why":"Supplies the evolving 2.4 μm galaxy luminosity function and its MCMC posterior chains, which are integrated to obtain the background.","marker":"Lake et al. (2018)"},{"why":"Supplies the mean galaxy spectral energy distribution used in the integral.","marker":"Lake & Wright (2016)"},{"why":"Provides the SED templates from which the mean galaxy SED is constructed.","marker":"Assef et al. (2010)"},{"why":"Defines the spectro-luminosity functional formalism that connects the luminosity function and mean SED to the background.","marker":"Lake et al. (2017)"},{"why":"Provides the WMAP9 cosmology used for distances, lookback times, and redshift conversions.","marker":"Hinshaw et al. (2013)"},{"why":"Defines the zCOSMOS survey whose deep multi-band photometry is used for the mean SED.","marker":"Lilly et al. (2009)"},{"why":"Supplies the zCOSMOS spectroscopic catalog used in the SED fitting.","marker":"Knobel et al. (2012)"},{"why":"Sets a blazar-based upper limit that the obtained background must be consistent with.","marker":"Mazin & Raue (2007)"},{"why":"A previous luminosity-function-based EBL estimate that the paper compares against.","marker":"Dominguez et al. (2011)"}],"fun_headline_variants":["Galaxies supply 9 kJy/sr of 3.4-micron cosmic background","Galaxies fill 3.4-micron cosmic background to 9 kJy/sr","WISE galaxy counts set 3.4-micron background at 9 kJy/sr","Galaxy census measures 3.4-micron background at 9 kJy/sr"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The calculation assumes that the adopted evolving Schechter luminosity-function model—with a fixed faint-end slope and a mean galaxy SED that does not change with redshift or luminosity—is the true description of the galaxy population, and a different but defensible version of the same model gives about 5 kJy/sr instead of 9.","fun_headline_variants_meta":{"raw":{"variants":["Galaxies supply 9 kJy/sr of 3.4-micron cosmic background","Galaxies fill 3.4-micron cosmic background to 9 kJy/sr","WISE galaxy counts set 3.4-micron background at 9 kJy/sr","Galaxy census measures 3.4-micron background at 9 kJy/sr"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001115,"raw_usage":{"total_tokens":4697,"prompt_tokens":1055,"completion_tokens":3642,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":671,"completion_tokens_details":{"reasoning_tokens":3542}},"tokens_in":671,"tokens_out":3642,"duration_ms":26097,"temperature":1.0,"reasoning_tokens":3542,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:34:00.517069+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A deep spectroscopic survey measuring the faint-end slope of the 2.4 μm luminosity function at $z \\approx 0.5$–$1$ to an uncertainty of $\\pm0.1$ would discriminate: if the slope is as steep as $\\alpha \\lesssim -1.9$, the predicted background falls to roughly $5\\,\\mathrm{kJy\\,sr^{-1}}$. Alternatively, a fully completeness-corrected W1 galaxy count at the faint end would test the model directly, since the paper's own comparison shows the preferred model over-predicting the faint counts.","supporting_citations":[{"cited_title":"$K$-corrections: an Examination of their Contribution to the Uncertainty of Luminosity Measurements","cited_arxiv_id":"1603.07299","evidence_quote":"Supplies the mean galaxy spectral energy distribution used in the integral."},{"cited_title":"E., Wright , E","cited_arxiv_id":null,"evidence_quote":"Defines the spectro-luminosity functional formalism that connects the luminosity function and mean SED to the background."},{"cited_title":"J., Iovino , A., et al","cited_arxiv_id":null,"evidence_quote":"Supplies the zCOSMOS spectroscopic catalog used in the SED fitting."}],"review_version":1}