{"id":"146396b8-b608-4b62-a1bf-8d0f67827715","arxiv_id":"2506.14873","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Stacked ASAS-SN images reach a median 3-sigma surface brightness limit of 26.1 mag per square arcsecond off the Galactic plane and recover 82 percent of known ultra-diffuse galaxies in the DES footprint.","lead":"This paper stacks seven years of ASAS-SN sky images into a new all-sky map that reaches faint surface brightness levels, revealing galactic shells, tidal features, and diffuse nebulae. It also recovers 82 percent of known ultra-diffuse galaxies in a test patch, suggesting the dataset could support a future all-sky search for faint dwarf galaxies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Flat-field gradients acknowledged in Section 4 are unquantified, so the 26.1 mag/arcsec^2 per-pixel limit is not demonstrated for extended sources, and the 82% recovery benchmark is not yet a blind validation.","rationale":"I read the paper as an honest proof-of-concept rather than a completed survey claim, and I agree with the conditional verdict. The central quantitative claims are the 26.1 mag/arcsec^2 depth and the 82% UDG recovery. Both depend on whether the stacked-image background is understood at the claimed level. The paper itself flags in Section 4 that flat fields are inaccurate at about the 1% level and create degree-scale gradients. Because the survey targets extended, low-surface-brightness features, those gradients are not a minor calibration detail; they are a potential systematic floor for exactly the science cases advertised. The per-pixel 3-sigma limit does not capture correlated noise, so the 26.1 value is not yet shown to be a usable extended-source sensitivity. The recovery test is useful but is not blind: it uses known SMUDGes positions plus visual inspection, and some of the 17 missed objects are explicitly attributed to flat-field edges. These limitations do not invalidate the paper, but they mean the headline numbers should be read as promising but not yet established. The concrete check I propose would measure the residual background covariance in blank fields, which would directly determine whether the systematics are comparable to the quoted limit. If that test passes, the central claims are much stronger; if it fails, the depth and recovery numbers need revision. This does not change the reader's conditional verdict, so I set verdict_should_be to UNCHANGED.","tokens_in":10843,"tokens_out":12063,"duration_ms":123797,"concrete_test":"Measure the residual background covariance directly: on roughly 20 blank-sky fields with |b| > 30°, select regions free of cirrus and bright stars, remove compact sources, and compute the two-point autocorrelation function (or power spectrum) of the background in each stacked image. Compare the variance at scales from 1 arcmin to 1 degree with the photon-noise expectation implied by the per-pixel 3-sigma = 26.1 mag/arcsec^2 limit. If the correlated large-scale variance exceeds the equivalent of a 3-sigma detection at 26.1 mag/arcsec^2 over a 16-arcsecond aperture, then the depth claim is not met for extended sources and the recovery benchmark should be re-evaluated with a position-dependent completeness test.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing assumption is that the stacked-image background is understood down to the quoted surface-brightness limits. The authors state in Section 4 that the existing flat fields are inaccurate at the ~1% level and produce degree-scale gradients in the stacks; this is acknowledged to make low-surface-brightness features near field borders hard to identify. The quoted median 3-sigma limit of 26.1 mag/arcsec^2 is a per-pixel photon-noise number, but for the extended sources this survey targets (cirrus, tidal tails, UDGs with Reff ≥ 16\"), the relevant noise is correlated large-scale background structure. A ~1% flat-field residual is a multiplicative error on the sky; its surface-brightness contribution is μ_sky + 5 mag/arcsec^2, which can fall within about a magnitude of the claimed limit for plausible ASAS-SN sky levels. The paper does not report the amplitude or angular scale of these residuals in the final stacks, so it is not established that the stacks are photon-noise-limited rather than systematics-limited. The 82% SMUDGes recovery rate is an encouraging empirical check, but it is measured with known positions and visual inspection, and the authors attribute some misses to flat-field edges; it therefore does not by itself validate the 26.1 limit or the future blind all-sky search.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This proof-of-concept paper describes the construction of stacked g-band images from roughly seven years of ASAS-SN data, covering the full sky with a median cumulative exposure time of 58.1 hours per field. The authors report a median 3-sigma surface brightness limit of 26.1 mag arcsec^-2 per 7.8-arcsecond pixel off the Galactic plane, illustrate the survey's ability to reveal Galactic nebulae, cirrus, and tidal features, and benchmark its performance by recovering 82% of ultra-diffuse galaxies from the SMUDGes catalog with effective radii of at least 16 arcseconds in the DES footprint. The paper concludes with a plan to release a public all-sky low-surface-brightness mosaic and to search for dwarf galaxies and cirrus.","tokens_in":11068,"tokens_out":6938,"duration_ms":65534,"significance":"If the claimed depth and recovery rate hold up, this dataset would be a unique community resource: no other survey currently combines all-sky coverage, roughly 26 mag arcsec^-2 surface brightness depth, and degree-scale angular reach. The paper has real strengths: calibration is tied to an external catalog (Refcat), the recovery benchmark uses the independent SMUDGes catalog, and the science figures convincingly show known low-surface-brightness structures. The main weaknesses are that the quoted depth is a per-pixel photon-noise number rather than a demonstrated extended-source sensitivity, and the UDG recovery test is not blind and reports no false-positive rate. The project is promising and within the scope of an astronomical instrumentation/methods journal, but the headline claims need additional support before they can be taken at face value.","major_comments":[{"comment":"The statement in Section 3.3 that an LMC-like object could be detected out to roughly 45 Mpc uses the nominal surface brightness limit without accounting for the correlated flat-field systematics discussed in Section 4. This distance estimate should be revisited once the background residual amplitude is measured, and the assumptions about angular size and aperture should be stated explicitly.","section":"Section 3.3, final paragraph"}],"minor_comments":[{"comment":"The abstract reports a median cumulative exposure time of 58.1 hours per field, but the surface brightness limits shown in Figure 1 are for the deepest single-camera stack, which contains on average only about 35% of the total exposure. Please clarify this distinction in the abstract and in the discussion of Figure 1 so that readers do not connect the 58.1-hour exposure to the 26.1 mag arcsec^-2 limit.","section":"Abstract and Section 2"},{"comment":"The caption states a surface brightness limit of 26.67 without giving units; please add mag arcsec^-2.","section":"Section 3.2, Figure 5 caption"},{"comment":"The horizontal axis label uses 'g (mag/arcsec2)' while the text refers to central surface brightness; please use a consistent notation such as mu_g and define it in the caption.","section":"Section 3.3, Figure 7"},{"comment":"The entries for Liu et al. 2023a and Liu et al. 2023b appear to refer to the same paper and DOI; if so, the duplicate citation should be corrected.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"This is a first paper in a series and the limitations are mostly acknowledged in Section 4, so the requested revisions are feasible within the scope of the manuscript. The main concern is that the abstract's headline numbers (26.1 mag arcsec^-2 and 82% recovery) go beyond what is currently demonstrated. I would be comfortable with a revised version that quantifies the flat-field systematics and adds an injection-recovery test with a false-positive rate. The paper fits the journal well and does not raise novelty or citation concerns beyond the duplicate reference noted in the minor comments."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does what a proof-of-concept should: it shows that stacking ~7 years of ASAS-SN g-band data produces images that reach a median 3-sigma surface brightness limit of 26.1 mag/arcsec2 per pixel off the Galactic plane, and that these images detect known UDGs, tidal features, and Galactic cirrus. That is a new data product, not a restatement of old results, and the all-sky coverage is a real gap-filler compared to DES or Legacy Survey based LSB work. The external benchmark against SMUDGes is a sensible choice, and the authors are transparent about the limits of their recovery test. Credit where due: they explicitly say the flat fields are only good to ~1%, that this produces degree-scale gradients, and that they plan a 2D polynomial fix. They also note the recovery rate comes partly from visual inspection with known positions. That honesty is what makes the paper readable.\n\nThe soft spot is the one the stress-test flags. The quoted 26.1 mag/arcsec2 is a per-pixel photon-noise number. For the extended sources this survey targets, correlated background structure from flat-field residuals is what matters, and the paper does not quantify the amplitude or angular scale of those residuals in the final stacks. A ~1% flat-field error on a typical sky level can contribute surface brightness within about a magnitude of the claimed limit, so it is not established that the stacks are photon-noise limited on the scales of UDGs or cirrus. The 82% recovery rate is encouraging but not a blind validation; it is measured against a catalog with known positions and includes visual inspection, with no false-positive rate. The authors themselves attribute several misses to flat-field edges. So the central claim of LSB sensitivity for extended sources is conditionally supported, not fully demonstrated.\n\nNone of this is fatal for a proof-of-concept. The paper is honest about what it does not yet know, and the next steps (combining cameras, better flat-fielding, a full public atlas) are clearly laid out. The right fix in review is to ask for a quantification of the flat-field residuals, a limit expressed on extended-source scales (e.g., 10x10 arcsec2 or a matched-aperture measurement), and a blind test or at least an estimate of the false-positive rate for the UDG recovery.\n\nThis deserves serious peer review. It is a solid methods paper with clear community value, and the flaws are in the details rather than in the core idea. I would not desk-reject it. If I worked on LSB galaxies or survey strategy, I would cite it as the first demonstration that ASAS-SN can do this. It is worth a reading group discussion precisely because the gap between per-pixel limits and extended-source sensitivity is a common trap.","headline":"A genuinely useful proof-of-concept for an all-sky LSB atlas; the per-pixel depth numbers are fine, but the extended-source sensitivity is not yet demonstrated because flat-field gradients are acknowledged but unquantified.","tokens_in":11698,"tokens_out":1760,"would_cite":true,"duration_ms":35485,"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":"Stacked ASAS-SN images reach a median 3-sigma g-band limit of 26.1 mag arcsec^-2 per pixel and recover 82 percent of ultra-diffuse galaxies with effective radius at least 16 arcseconds in the DES footprint.","keywords":["low surface brightness","ASAS-SN","image stacking","ultra-diffuse galaxies","galactic cirrus","tidal features","all-sky survey","surface brightness limits"],"falsifier":"Compute the residual background rms of the final stacked images after applying a 2D polynomial flat-field correction (as described in Section 4) on fields away from the Galactic plane; if the residual rms is larger than the nominal 3-$\\sigma$ surface brightness limit of 26.1 mag $arcsec^{-2}$, the effective depth for extended sources is below the quoted value. Also cross-check detected ultra-diffuse galaxies against higher-resolution DES or Legacy images to confirm that sources near ASAS-SN field edges are real and not flat-field artifacts.","tokens_in":10623,"feed_emoji":"🌌","tokens_out":9855,"duration_ms":81091,"temperature":0.7,"pith_summary":"This paper claims that archival images from the ASAS-SN supernova survey, repurposed and stacked, deliver a deep all-sky view of low-surface-brightness structure. Roughly seven years of g-band data, with a median cumulative exposure of 58.1 hours per field, reach a median 3-$\\sigma$ surface brightness limit of 26.1 mag $arcsec^{-2}$ per pixel away from the Galactic plane, several magnitudes deeper than single-night images. The authors demonstrate the result by imaging degree-spanning Milky Way nebulae, tidal tails around NGC 3628, shells around NGC 474, and by recovering 82 percent of catalogued ultra-diffuse galaxies with effective radius of at least 16 arcseconds in the DES footprint. If this holds, it would give astronomers a new all-sky atlas of the faint sky, including dwarf galaxies, Galactic cirrus, and galaxy-halo features, that no dedicated low-surface-brightness telescope has produced.","feed_headline":"26.1 mag/arcsec^2: ASAS-SN stacks map the faint sky","feed_subtitle":"Repurposed supernova-hunting images recover 82% of known ultra-diffuse galaxies and reveal Milky Way nebulae.","key_machinery":"The stacker scales each exposure by its median sky value before coadding; because the variance in blank sky is proportional to the sky level, the sum of I_i/S_i is an inverse-variance-weighted mean equivalent to I_i/$sigma_i^{2}$, the optimal way to combine noisy backgrounds. Iterative pixel-by-pixel 3-$\\sigma$ outlier rejection cleans cosmic rays and moving objects, and the top 10 percent of images with the highest sky backgrounds are discarded. For the galaxy search, stars are first masked or PSF-subtracted using Gaia proper-motion and parallax cuts plus Gaussian and Moffat profile fits, and sources are then found with SourceExtractor using a 3x3-pixel kernel, supplemented by visual inspection.","core_discovery":"This paper claims that stacking about seven years of g-band ASAS-SN images, taken with 7.8-arcsecond pixels and median exposure times near 58 hours per field, reaches a median 3-$\\sigma$ surface brightness limit of 26.1 mag $arcsec^{-2}$ per pixel for fields more than 20 degrees from the Galactic plane. It recovers 37 of 92 SMUDGes ultra-diffuse galaxies automatically and 38 more by visual inspection, for an 82 percent recovery rate among UDGs with effective radius at least 16 arcseconds in the DES footprint. The same stacks reveal Milky Way nebulae spanning several degrees, Galactic cirrus, and tidal features around NGC 3628 and NGC 474, which the authors present as proof that a time-domain survey can double as a wide-area low-surface-brightness survey.","pith_inferences":["The paper leaves implicit that, because the flat-field gradients limit reliability near field edges, the 82 percent recovery rate may be a conservative estimate for large galaxies away from edges; applying the 2D polynomial correction described in Section 4 before the next search should push the usable area deeper.","Combining all cameras per field, rather than only the best single camera, would recover the full cumulative exposure time and could deepen the stacks by roughly the 0.4 mag implied by the exposure-time comparison, opening up slightly fainter and more distant dwarfs.","An all-sky cirrus map built from these stacks could serve as a foreground mask for upcoming deep surveys, separating Galactic dust emission from genuine extragalactic low-surface-brightness features.","Because the ASAS-SN point-spread function is well matched to ultraviolet and infrared all-sky surveys, cross-correlating the same stacks at other wavelengths could help separate stellar streams from cirrus, a testable extension the paper identifies as a natural next step."],"forward_implications":["Seventy-five percent of all ASAS-SN fields, and 97 percent of fields more than 20 degrees from the Galactic plane, reach surface brightness limits of 24 mag arcsec^-2 or deeper, the formal threshold of the low-surface-brightness regime.","An LMC-like dwarf galaxy could be detected out to roughly 45 Mpc and an SMC-like dwarf to roughly 20 Mpc, so an all-sky search with these stacks can find nearby dwarf galaxies that previous surveys missed.","The atlas resolves tidal tails and shells around nearby galaxies, enabling surface brightness profiles of galaxy halos to be measured over the whole sky.","The same data can be used to produce an all-sky Galactic cirrus map, helping to mask regions of diffuse dust contamination for deeper surveys.","A public, full-sky g-band low-surface-brightness mosaic is the planned end product, giving the community access to these images."],"supporting_citations":[{"why":"Describes the ASAS-SN survey whose g-band images are stacked to make the low-surface-brightness atlas.","marker":"Shappee et al. 2014"},{"why":"Documents the ASAS-SN data pipeline and survey operations that produce the nightly images used here.","marker":"Kochanek et al. 2017"},{"why":"Provides the SMUDGes catalog of ultra-diffuse galaxies used to measure the 82 percent recovery rate.","marker":"Zaritsky et al. 2023"},{"why":"Supplies the Refcat calibration catalog used to photometrically calibrate the stacked images.","marker":"Tonry et al. 2018"},{"why":"Provides the proper-motion and parallax measurements used to identify and subtract foreground stars.","marker":"Gaia Collaboration et al. 2023"},{"why":"SourceExtractor is the source-detection algorithm used to automatically find low-surface-brightness galaxies.","marker":"Bertin & Arnouts 1996"},{"why":"Defines the DES footprint inside which the ultra-diffuse galaxy recovery test is performed.","marker":"The Dark Energy Survey Collaboration 2005"}],"fun_headline_variants":["Stacked ASAS-SN images hit 26.1 mag/arcsec²","82% recovery of ultra-diffuse galaxies from stacked data","Reusing SN-hunting data for all-sky faint structure survey"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the stacked images can be trusted at low surface brightness across each field even though the calibration frames used to correct the camera's uneven response are only accurate to about one percent and create wide, smooth gradients across the image; if those gradients add correlated background structure at the level of the quoted limits, the true sensitivity for extended objects is lower and the recovery rate is optimistic.","fun_headline_variants_meta":{"raw":{"variants":["Stacked ASAS-SN images hit 26.1 mag/arcsec²","82% recovery of ultra-diffuse galaxies from stacked data","Reusing SN-hunting data for all-sky faint structure survey"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001075,"raw_usage":{"total_tokens":4497,"prompt_tokens":937,"completion_tokens":3560,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":553,"completion_tokens_details":{"reasoning_tokens":3498}},"tokens_in":553,"tokens_out":3560,"duration_ms":26297,"temperature":1.0,"reasoning_tokens":3498,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T00:09:45.992460+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the residual background rms of the final stacked images after applying a 2D polynomial flat-field correction (as described in Section 4) on fields away from the Galactic plane; if the residual rms is larger than the nominal 3-$\\sigma$ surface brightness limit of 26.1 mag $arcsec^{-2}$, the effective depth for extended sources is below the quoted value. Also cross-check detected ultra-diffuse galaxies against higher-resolution DES or Legacy images to confirm that sources near ASAS-SN field edges are real and not flat-field artifacts.","supporting_citations":[],"review_version":1}