{"id":"9fc993df-5d87-4a50-9366-22fb1a875d57","arxiv_id":"2608.05531","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"All-sky maps of stellar mass and star formation rate in the local universe, binned by telescope field-of-view and distance, provide optimized pointing priors for transient surveys.","lead":"Astronomers built all-sky maps of galaxy stellar mass and star formation rate out to 200 million light-years, at resolutions matched to real telescopes, to show where local cosmic explosions are most likely to be found. The maps are meant to help survey planners point their telescopes at the richest regions instead of scanning empty sky.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Uncorrected sky-coverage incompleteness in REGALADE is the load-bearing risk: the anisotropy maps may encode survey selection, not intrinsic mass/SFR density.","rationale":"I read the paper in good faith. The data product is useful and the authors are transparent about many limitations. The central claim is conditional on the maps representing intrinsic density fields. The reader's weakest assumption identifies exactly the issue I find most load-bearing: the lack of a quantitative completeness correction. I considered other potential concerns—the dwarf-galaxy incompleteness estimate (§4.1), the angular power spectrum estimator (§4.2), and the qualitative CCSN validation (§4.3)—but none is as central: dwarf galaxies contribute a small mass fraction, the power spectra are not the main deliverable, and the CCSN comparison is explicitly framed as qualitative. The completeness and SFR-extrapolation issues directly affect the map values that a survey planner would act on. I do not see an internally inconsistent argument or evidence of fabrication; the concern is about unquantified selection effects, which the authors themselves flag. My recommendation is therefore to keep the reader's CONDITIONAL verdict: the paper should be published with the condition that either a completeness correction or an explicit demonstration that completeness does not drive the anisotropy be added. I agree with the reader's identification of the weakest assumption, so agreement_with_reader = agree.","tokens_in":20565,"tokens_out":3662,"duration_ms":34254,"concrete_test":"Build a REGALADE completeness map for D<200 Mpc by computing, in each HEALPix pixel at θ=1.83°, the expected recovery fraction from the union of the component survey footprints and depth limits (or, as a proxy, the ratio of galaxy counts in a nearby shell to counts in a more distant shell). Cross-correlate this map with the stellar mass and SFR maps; if the cross-correlation is comparable to the maps' autocorrelation, the anisotropy is selection-dominated. A complementary check: regenerate the maps using only galaxies with spectroscopic redshifts from a volume-limited, all-sky sample (e.g., 2MRS or a 2MASS-based sample) and compare the location of the top decile of cells. If the top pointing directions shift by more than one or two cells, the roadmap is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that the mapped stellar mass/SFR anisotropies can guide transient pointing—requires that the observed large-scale structure is astrophysical rather than an imprint of the survey footprint. In §2.2 the authors state that REGALADE has 'non-uniform sky coverage and varying detection depths' and 'severe incompleteness near the Galactic plane and bulge region.' The maps in Figs. 3 and 4 are raw HEALPix sums over the catalog with no completeness weighting or selection-function correction. Masking ±10° in latitude for the power spectra (Fig. 5) removes only the most obvious plane artifact; it does not remove the large-scale footprints of the constituent surveys (SDSS, Pan-STARRS, DES, 2MASS, WISE) that REGALADE merges, and the full-sky maps themselves retain the plane region. If the angular completeness pattern correlates with the claimed overdensities, the 'roadmap' partly encodes which surveys happened to cover which sky. The authors acknowledge the limitation but do not quantify its angular imprint, and the CCSN comparison in §4.3 is explicitly qualitative and cannot distinguish selection from signal. A second layer is the SFR extrapolation: the CatBoost model is trained on GSWLC-2 (SDSS footprint) and applied to 325,807 galaxies all-sky (§2.3.3) with no test for systematic SFR bias across different photometric systems. Both layers bear directly on the survey-planning use case, so this is the single most load-bearing concern.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs all-sky maps of galaxy stellar mass and star formation rate (SFR) within luminosity-distance thresholds of 30-200 Mpc, at angular resolutions matched to the fields of view of Mephisto, ZTF, and Einstein Probe. The maps are built from the REGALADE galaxy catalog, with SFRs for 325,807 galaxies predicted by a CatBoost model trained on GSWLC-2 labels and applied all-sky. The authors find that the stellar-mass and SFR maps share their main anisotropic structures, that these structures become weaker as the distance threshold increases, and that the angular power spectra and fluctuations decrease with distance. They also compare the SFR maps qualitatively with the sky distribution of core-collapse supernovae from BSN and TNS and report broad consistency in several prominent structures. The stated goal is to provide a 'roadmap' for transient surveys with medium fields of view.","tokens_in":20800,"tokens_out":3291,"duration_ms":32649,"significance":"If the maps faithfully represent the intrinsic nearby stellar-mass and SFR density fields, they would provide a practical, publicly available pointing prior for medium-field transient surveys and a flexible framework for matching angular resolution to a telescope's field of view. A clear strength is that the full dataset, training samples, and codes are released on Zenodo, and the authors are explicit that the supernova comparison is qualitative. However, the central survey-planning claim rests on the assumption that the observed large-scale anisotropies are astrophysical rather than imprints of the heterogeneous depth and sky coverage of the input catalogs, and on the assumption that a GSWLC-2-trained SFR model generalizes all-sky. These assumptions are acknowledged but not quantitatively tested, which is the main barrier to accepting the maps as a roadmap.","major_comments":[{"comment":"The maps are raw HEALPix sums over REGALADE with no completeness or selection-function correction, even though the paper states in §2.2 that the catalog has 'non-uniform sky coverage and varying detection depths' and 'severe incompleteness near the Galactic plane and bulge region.' If the angular completeness pattern of the constituent surveys correlates with the claimed overdensities, the roadmap partly encodes which surveys covered which sky. Because this is load-bearing for the central claim, the authors should either apply or construct completeness weights for REGALADE, or demonstrate that the main structures in Figs. 3 and 4 survive when the analysis is restricted to regions with uniform coverage or repeated on individual constituent surveys. A quantitative comparison of the anisotropy maps with a survey-footprint map would also address the concern directly.","section":"§2.2 and Figs. 3-4"},{"comment":"The CatBoost SFR model is trained on GSWLC-2, whose training set is dominated by SDSS photometry, and is then applied to the full all-sky REGALADE sample. The features include g, r, z, W1, W2, redshift, and stellar mass, but the photometric systems supplying g, r, and z vary across the sky between Legacy Surveys, DELVE, and Pan-STARRS. The paper reports no test for systematic SFR bias as a function of sky region or photometric system. A concrete test would be to predict SFRs for a held-out all-sky sample with known SFRs outside the SDSS footprint, or to compare the feature distributions of the target sample with the training distribution; without such a test, the SFR maps may contain large-scale systematics that mimic or suppress the anisotropy signal.","section":"§2.3.3"},{"comment":"The angular power spectra in Fig. 5 are presented without error bars or any estimate of sample variance, mask-induced mode coupling, or completeness uncertainty. The main trend claim—that the power spectra decrease rapidly with increasing distance threshold—would be more convincing with at least jackknife or bootstrap uncertainties over independent sky regions, and with a discussion of how the ±10° Galactic-plane mask and the catalog completeness pattern affect the measured power at low and high multipoles. Without this, the reader cannot tell whether the reported decline to 'zero' at 180 Mpc is significant or an artifact of the mask and selection function.","section":"§4.2 and Fig. 5"}],"minor_comments":[{"comment":"The color-bar labels 'log(M/M)' and 'log(SFR/M yr^-1)' are missing the solar-mass symbol and the proper superscript formatting; as printed they are easy to misread.","section":"Figs. 3-4"},{"comment":"The x-axis label 'line scale' is ambiguous; it should be labeled as multipole ℓ or as angular scale in degrees, and the panels would benefit from uncertainty bands.","section":"Fig. 5"},{"comment":"The training and blind-test RMSE values are both reported as 0.306 to three decimal places, which is suspicious for a CatBoost model; please clarify whether this reflects rounding or a metric that is dominated by the most common galaxy populations, and report the blind-test metrics separately with more precision.","section":"§2.3.2 and Table 1"},{"comment":"The authors correctly label the CCSN comparison as qualitative and list the selection biases of the input catalogs; a simple quantitative summary, such as a rank correlation between the SFR map and CCSN surface density after basic distance cuts, would strengthen the practical roadmap claim without overstating the selection-function control.","section":"§4.3"}],"recommendation":"major_revision","confidential_remarks":"The paper is timely and provides a useful public product, but the two major concerns—uncorrected sky-coverage incompleteness and all-sky application of an SDSS-trained SFR model—are directly load-bearing for the roadmap claim. These are fixable within the manuscript's scope by adding robustness tests or completeness-weighting demonstrations, so I recommend major revision rather than rejection. The qualitative CCSN comparison is appropriately hedged and should not, by itself, block publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Useful data product, honestly packaged, but the sky-coverage incompleteness in the input catalog is a real soft spot that should stop a survey planner from treating the maps as a final answer without further checks. The genuinely new piece here is the all-sky SFR layer: they take REGALADE, predict SFRs for ~325k galaxies with a CatBoost model trained on GSWLC-2, and produce HEALPix maps at 30–200 Mpc for FoVs matched to Mephisto, ZTF, and EP. The maps themselves, the angular power spectra, and the fluctuation curves are all new analyses, and the Zenodo release with maps and code makes the product easy to reuse. The machine-learning performance is solid: RMSE 0.306 on a blind test, slightly better than the Z24 reference, and the residuals look Gaussian. They also deserve credit for the dwarf-galaxy completeness estimate—rough, but a sensible back-of-the-envelope—and for being explicit that the CCSN comparison is qualitative. The soft spots are the ones the authors themselves flag but do not quantify. REGALADE has non-uniform sky coverage and varying depth, and the maps are raw HEALPix sums with no completeness correction. Masking ±10° in latitude for the power spectra removes the most obvious plane artifact but not the footprints of the constituent surveys (SDSS, Pan-STARRS, DES, 2MASS, WISE). If those footprints correlate with the overdensities they highlight, the 'roadmap' partially encodes which surveys happened to cover which part of the sky. For the 50 Mpc maps this is probably a minor effect because the dominant structures (Virgo, Fornax, the local sheet) are real and well known, but for a quantitative survey optimizer it is load-bearing. The second concern is the SFR extrapolation: the CatBoost model is trained on SDSS-footprint photometry and applied all-sky, with no test for systematic offsets between photometric systems. That is an easy fix—split the validation by survey—but it is not in the paper. The power spectra also have no error bars or mask treatment, which limits their use beyond the qualitative trend. None of this kills the paper. The central anisotropy-to-isotropy trend is robust to these issues, and the authors are careful not to overclaim the CCSN comparison. The paper would benefit from a completeness-weighting exercise and a cross-survey SFR validation, both of which are straightforward. As it stands, it is a useful data product for transient survey strategy, especially for medium-FoV telescopes like Mephisto. I would send it to a serious referee. The flaws are addressable, the data are public, and the product fills a practical niche that no other catalog provides at this resolution.","headline":"Useful, honest data product for survey planners, but the uncorrected sky-coverage incompleteness in REGALADE is a real soft spot that should be addressed before the maps are used as a quantitative roadmap.","tokens_in":21376,"tokens_out":2676,"would_cite":false,"duration_ms":24483,"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":"Sky maps of galaxy mass and star formation point to transient hotspots","keywords":["Star formation","Stellar masses","Sky surveys","Transient detection","Time domain astronomy","Angular power spectrum","Galaxy catalog","Core-collapse supernovae"],"falsifier":"Compare the maps against a volume-limited galaxy sample built from a different all-sky catalog with independent distance and SFR estimates; if the dominant anisotropy changes when the catalog changes, the maps partly encode selection. A stronger test would run a uniform wide-field transient survey within 200 Mpc and check that observed events per pixel follow the maps after accounting for the survey's own selection function, with a mismatch tracking the catalog's known gaps (Galactic plane, LMC/SMC directions) falsifying the roadmap.","tokens_in":20283,"feed_emoji":"🔭","tokens_out":5549,"duration_ms":46194,"temperature":0.7,"pith_summary":"The paper argues that the anisotropic sky distributions of galaxy stellar mass and star formation rate out to about 200 Mpc can serve as a practical roadmap for transient surveys, especially for telescopes with fields of view of a few to tens of square degrees. It builds all-sky maps from a large galaxy catalog, predicts star formation rates for galaxies lacking them, and bins the total stellar mass and SFR into pixels matched to telescope fields of view and distance thresholds. The maps show strong large-scale structures that fade toward isotropy as the distance limit grows, and the angular power spectra and fluctuations of both quantities drop rapidly with distance. A qualitative comparison with core-collapse supernovae finds that confirmed explosions cluster in the same prominent structures as the SFR maps. If the maps are correct, pointing at their overdensities should raise the probability of discovering local transients compared with blind scanning.","feed_headline":"Sky maps of galaxy mass and star formation point to transient hotspots","feed_subtitle":"Pointing at galaxy overdensities within 200 Mpc should beat blind scanning for finding nearby explosions.","key_machinery":"The load-bearing mechanism is the mapping of total stellar mass and SFR per equal-area HEALPix pixel, with pixel size set by telescope field of view and the galaxy sample cut by luminosity distance. Transient rates are connected to the maps through the two-component 'A+B' model, Rate = A (M*/$10^{10}$ Msun) + B (SFR/10 Msun/yr), so events with short delays follow the SFR field and delayed events follow a mix of stellar mass and SFR. The SFR field itself rests on a CatBoost regressor, and the angular power spectrum and fluctuation statistic quantify how fast anisotropy dies with distance.","core_discovery":"Using the REGALADE galaxy catalog, the paper constructs all-sky maps of total stellar mass and total star formation rate for galaxies within 200 Mpc, at angular resolutions of 1.83 degrees, 7.33 degrees, and 58.6 degrees matched to the fields of view of Mephisto, ZTF, and Einstein Probe, and for distance thresholds from 30 to 200 Mpc. Star formation rates for 325,807 galaxies are predicted with a CatBoost model trained on GSWLC-2 measurements, giving a complete SFR census of 364,148 galaxies. The central claim is that these distributions, which the paper calls a roadmap for transient hunters, trace the expected rate densities of local transients: stellar-mass-linked events such as Type Ia supernovae and compact-object mergers track the stellar mass maps, while SFR-linked events such as core-collapse supernovae and long gamma-ray bursts track the SFR maps. The maps become increasingly isotropic with distance, and the angular power spectra and relative fluctuations of both fields decline rapidly, quantifying the zone within which anisotropy actually matters for survey design.","pith_inferences":["A direct test would be to simulate a survey with known cadence and sensitivity, draw transient rates from the maps, and check that the pointing strategy maximizing recovered events matches the map overdensities; the paper does not run this simulation.","The strongest systematic risk not tested in the paper is that the SFR model, trained on SDSS-footprint galaxies, may carry photometric-system biases when applied all-sky; comparing predicted SFRs for galaxies with independent UV/IR SFR estimates outside the SDSS footprint would settle this.","The maps could be combined with transient rate densities per unit stellar mass and SFR to forecast relative event rates for different transient classes, not just spatial patterns; the paper stops at spatial correlation.","Near the Galactic plane and the LMC/SMC directions, catalog incompleteness could suppress apparent densities, so surveys should treat those regions as lower-confidence zones."],"forward_implications":["A medium-field telescope that points its exposures at the overdense pixels in these maps should discover more local transients per pointing than one that scans blindly.","The published dataset lets any survey choose its own pixel size and distance cut, so the roadmap generalizes to fields of view beyond the three example telescopes.","Because the stellar mass and SFR maps differ in their angular power spectra, mass-linked and SFR-linked transients should show measurably different sky clustering in the local universe.","The rapid decline of anisotropy beyond roughly 100-180 Mpc means survey optimization from these maps matters mainly for nearby targets; beyond that, blind scanning loses little.","The maps can serve as intrinsic prior maps onto which a future uniform survey's selection function can be convolved for quantitative transient-rate studies."],"supporting_citations":[{"why":"Supplies the REGALADE all-sky galaxy catalog with stellar masses, distances, and photometry that the maps are built from.","marker":"Tranin et al. 2026"},{"why":"Provides the GSWLC-2 SFR measurements used both as machine-learning training labels and as direct SFR values for a subset of the sample.","marker":"Salim et al. 2018"},{"why":"Establishes the CatBoost methodology and the benchmark whose RMSE the paper's SFR predictor is compared against.","marker":"Zeraatgari et al. 2024"},{"why":"Defines the HEALPix equal-area pixelization used to bin stellar mass and SFR on the celestial sphere.","marker":"Górski et al. 2005"},{"why":"Formalizes the two-component A+B model linking transient rates to a delayed stellar-mass term and a prompt SFR term.","marker":"Scannapieco & Bildsten 2005"},{"why":"Supports the premise that short-lived massive stars and prompt core-collapse transients trace ongoing star formation.","marker":"Kennicutt & Evans 2012"},{"why":"Provides the galaxy stellar mass function used to argue that missing dwarf galaxies contribute only about 1.78 percent of stellar mass.","marker":"Baldry et al. 2012"},{"why":"Sources one of the two supernova compilations whose core-collapse positions are compared qualitatively with the SFR maps.","marker":"Gal-Yam et al. 2013"}],"fun_headline_variants":["Transient hunters get a roadmap from local galaxy maps","Galaxy maps show where nearby supernovae are likeliest","Anisotropy maps guide searches for local transients","Local universe maps spotlight prime transient locations","Local galaxy maps chart transient hunting grounds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The maps assume that uneven sky coverage and depth in the underlying galaxy catalog do not create the large-scale patterns, so those patterns reflect real matter and star formation rather than survey selection.","fun_headline_variants_meta":{"raw":{"variants":["Transient hunters get a roadmap from local galaxy maps","Galaxy maps show where nearby supernovae are likeliest","Anisotropy maps guide searches for local transients","Local universe maps spotlight prime transient locations","Local galaxy maps chart transient hunting grounds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000816,"raw_usage":{"total_tokens":3621,"prompt_tokens":1037,"completion_tokens":2584,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":653,"completion_tokens_details":{"reasoning_tokens":2511}},"tokens_in":653,"tokens_out":2584,"duration_ms":15651,"temperature":1.0,"reasoning_tokens":2511,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T11:22:54.935796+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the maps against a volume-limited galaxy sample built from a different all-sky catalog with independent distance and SFR estimates; if the dominant anisotropy changes when the catalog changes, the maps partly encode selection. A stronger test would run a uniform wide-field transient survey within 200 Mpc and check that observed events per pixel follow the maps after accounting for the survey's own selection function, with a mismatch tracking the catalog's known gaps (Galactic plane, LMC/SMC directions) falsifying the roadmap.","supporting_citations":[{"cited_title":"K., Driver, S","cited_arxiv_id":null,"evidence_quote":"Provides the galaxy stellar mass function used to argue that missing dwarf galaxies contribute only about 1.78 percent of stellar mass."}],"review_version":1}