{"id":"89e575eb-7984-47a2-a6da-14e7f96e3618","arxiv_id":"2501.10623","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"First full-scale (single-antenna plus aperture-synthesis) diffuse Galactic Faraday rotation maps at 3' resolution, with stated caveats about narrow-band RM reliability.","lead":"The authors combined single-antenna and aperture-synthesis radio data to create the first diffuse Milky Way Faraday rotation maps that span all angular scales down to 3 arcminutes. The maps reveal magnetic-field structures across scales, but the 35 MHz bandwidth means individual rotation-measure values need cautious interpretation.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'all spatial scales' claim rests on a Gaussian-beam deconvolution and feathering whose only validation is a partly circular simulation that excludes known artifacts; a wrong beam model or underestimated overlap-region noise would bias every RM.","rationale":"The reader's weakest_assumption identifies the same load-bearing premise: the GMIMS-HBN visibilities after Gaussian deconvolution, and the feathering across the missing intermediate uv spacings, must accurately represent the true short-spacing sky. My review sharpens this into two specific technical weaknesses. First, the validation simulation is partly circular because it uses the same Gaussian beam model to create the mock single-antenna data that it later deconvolves; it therefore cannot detect errors in the beam model itself. Second, the simulation excludes the bright-source and artifact cases that the real data are known to contain, and the paper provides no per-pixel error map or quantitative comparison of the combined Stokes Q/U to the L10 maps that used Effelsberg 100 m data to bridge the gap. These are not fatal objections: the paper is explicitly a prototype experiment, the authors transparently discuss the limitations, and the power-spectrum comparison in Figure 14 gives some supporting evidence that the overlap region is not wildly discrepant. However, because the 'all spatial scales' claim depends on the large-scale Q/U being correct, and because a common-mode systematic in Q/U would propagate directly into every derived RM, conditional acceptance is appropriate pending the cross-validation test described above. I agree with the reader's overall assessment and verdict.","tokens_in":26621,"tokens_out":4419,"duration_ms":52315,"concrete_test":"Compute the difference between the new full-scale Stokes Q/U mosaics and the L10 combined Q/U mosaics (which included Effelsberg 100 m data) over the common sky, after matching resolution and excluding the flagged artifact regions; form the power spectrum of the difference. If the residual power in the 8-17 m baseline range (angular scales roughly 0.7-1.5 deg) is consistent with the 0.008 K simulation residual and with expected noise from the DRAO ST maps, the bridge is validated. In addition, rerun the Section 3.4.2 simulation with the beam HWHM perturbed by +/-10% and with a bright-source leakage term included; if the resulting Q/U differences exceed the 0.008 K level, the claimed all-scale coverage is not robust.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim requires the combined Stokes Q/U maps to represent the true sky on all scales. In Section 3.1(iii), GMIMS-HBN visibilities are divided by the Fourier transform of a Gaussian beam model with 9 m HWHM, and the resulting data are used out to 17.144 m in the feathering of Section 3.2 even though the single-antenna data are only described as reliable out to ~9 m. Beyond ~15 m the beam transform approaches zero, so the division amplifies noise; Figure 5 shows the GMIMS-HBN/ST visibility ratios peaking between 1 and 1.5 with a wide spread, indicating that the single-antenna data are not cleanly matching the interferometer in the overlap. The only validation, Section 3.4.2, is partly circular: it generates simulated single-antenna data by convolving the true image with the same Gaussian beam model that is later deconvolved, so it cannot detect error in the beam model itself. It also excludes bright sources and known artifacts, which the authors acknowledge in Section 4.1 (Cygnus X leakage, Cas A and W3 rings, the hook artifact at l=120 deg), and it reports a single residual realization rather than a per-pixel error map. If the beam model is imperfect or the overlap-region noise is underestimated, the degree-scale Stokes Q/U are systematically biased; because all four frequency channels share the same processing, the RM maps inherit a common-mode systematic error, not merely random noise. Without a quantitative check against data that actually bridge the uv gap (the L10 combined maps included Effelsberg 100 m data), this load-bearing premise remains unverified.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a proof-of-concept combination of single-antenna (GMIMS-HBN) and aperture-synthesis (CGPS Synthesis Telescope) Stokes Q/U data in four 7.5 MHz channels centred near 1420 MHz. The combination uses Gaussian-beam deconvolution in the uv plane, feathering of the two datasets, and mosaicking, after which per-pixel rotation measures are computed from linear fits of polarization angle against wavelength squared over the 35 MHz bandwidth. The resulting RM and PI maps cover the CGPS longitude range, are compared with ST-only, GMIMS-HBN-only, and broadband Faraday-depth data, and are interpreted in three regions (Sh2-216, IC 443, and the l = 173 deg H II complex). The authors state clear caveats about the narrow bandwidth, the impossibility of detecting Faraday complexity, a uv-coverage gap, and several known artifacts. The central claim is that this is the first full-scale polarization dataset combining single-antenna and aperture-synthesis data across multiple frequency channels, enabling diffuse-emission Faraday rotation studies of all spatial scales down to 3 arcmin.","tokens_in":26987,"tokens_out":4168,"duration_ms":46573,"significance":"If the method is sound, this is a useful pioneering data product: it is the first demonstration that diffuse Galactic RM information can be obtained from combined single-antenna and aperture-synthesis data across separate frequency channels, and it highlights a path for future broadband surveys (e.g., CHIME/DRAGONS, PEGASUS/POSSUM). The processing is transparently documented, the data products are public, and the comparison with L10, Dwingeloo, and compact-source RMs provides a degree of external grounding. The main caveat is that the short-spacing correction rests on a Gaussian beam model and a feathering procedure whose validation is partly circular and excludes known artifacts; because all four frequency channels are processed identically, any error there becomes a common-mode systematic in the RM maps. The paper's own discussion of the narrow-bandwidth lever-arm problem also shows that the numerical RM values are not simply the Faraday depths of the lines of sight. For these reasons the significance is real but conditional on the additional uncertainty quantification described in the major comments.","major_comments":[{"comment":"The simulation validates the feathering pipeline but not the Gaussian beam model that is the crux of the short-spacing correction. The simulated single-antenna image is generated by filtering the true image with the same 9 m HWHM Gaussian that is later deconvolved via Eq. (3), so any error in the beam model is invisible to the test. The reported 6% Stokes U error therefore applies only to the uv-coverage gap, not to beam-model systematics. Because all four frequency channels are processed identically, a beam-model error would introduce a common-mode bias in Stokes Q and U and hence in RM, rather than merely extra random noise. The authors should either validate the Gaussian beam model with an independent measurement or estimate the sensitivity of the final Q, U, and RM maps to plausible beam-model variations.","section":"§3.4.2"},{"comment":"The GMIMS-HBN data are used in the feathering out to 17.144 m although the data are described as reliable only to roughly 9 m. Figure 4b shows the deconvolved single-antenna visibilities diverging from the ST beyond about 15 m, and Figure 5 shows ratio distributions peaking between 1 and 1.5 with a wide spread, so the overlap region is noisy and not perfectly matched. The manuscript interprets these as tolerable, but no quantitative estimate is given for how the overlap-region mismatch propagates into RM uncertainties; the simulation in §3.4.2 injects noise only at the maximum baseline and reports one residual realization. A per-pixel or per-spatial-scale uncertainty map for Stokes Q, U, and RM, or at least a plausible upper bound on the systematic RM bias from this region, is needed to support the 'all spatial scales' claim.","section":"§3.1(iii) and §3.2"},{"comment":"The comparison between the narrowband RMs and the broadband GMIMS-HBN peak Faraday depths shows a large systematic offset (note the different vertical scales in panels b and c), and the ST-only diffuse RMs trace a different large-scale longitude pattern than the combined or single-antenna RMs. The paper attributes this to an inadequate lever arm and beam depolarization, but the consequence is that the derived RM values are not simply 'the' Faraday depth of each line of sight. Since the abstract and §6 nevertheless present the RM maps as the central product, the text should state more explicitly which quantitative statements (morphology, gradients, sign changes) are robust to this systematic effect and which are not. This is particularly important because §4.2 already states that Faraday complexity cannot be detected and that the RM values should be interpreted with caution.","section":"§5.3, Figure 21"},{"comment":"The RM maps are masked only by a PI threshold; no per-pixel uncertainties or fit-quality maps are provided with the released data. The sample fits in Figure 15 show RM uncertainties of tens of rad m^-2, but there is no way for a reader to assess which map features are significant. Since the FITS data products are publicly released, the authors should provide uncertainty maps (from the linear-fit covariance, including the correlated systematic component from the single-antenna processing) together with the RM, PI, Q, and U maps.","section":"§4.2, §5.2, Data Availability"}],"minor_comments":[{"comment":"The title contains an intra-word space in 'F araday'; this should be corrected to 'Faraday'.","section":"Title"},{"comment":"The caption refers to a '(green)' symbol for the CGPS point sources, but no green symbol is identified in the visible legend; please clarify the legend or the caption.","section":"Figure 21 caption"},{"comment":"The manuscript states that the initial instrumental polarization was estimated at 3% of total intensity and later reports 0.3% remaining leakage; a brief sentence explaining whether 0.3% refers to the residual after the iterative correction would remove ambiguity.","section":"§2.2"},{"comment":"The statement that feathering boundaries were chosen as 'integer multiples of the spacing between adjacent elliptical tracks' would be clearer if the relevant spacing in meters were stated explicitly, alongside the values 8.572 m and 17.144 m.","section":"§3.2"},{"comment":"The hook-shaped artifact at l = 120 deg, the rings around Cas A and W3, and the Cygnus X leakage are mentioned in the text, but the RM maps themselves do not mark these regions; a few additional contours or shaded regions on the figures would help readers avoid misinterpreting artifacts as real structures.","section":"§4.1 and §4.2"}],"recommendation":"major_revision","confidential_remarks":"This is a data-processing and method-demonstration paper rather than a discovery paper, and it is suitable for a journal that accepts such contributions. The authors are transparent about many limitations, which is commendable. The central issue is that the short-spacing correction, which is load-bearing for the 'all spatial scales' claim, is validated only by a partly circular simulation and the overlap-region noise is not propagated into RM uncertainties. I do not see this as an unfixable flaw: the requested uncertainty maps and beam-model sensitivity tests are within the scope of a revision. I recommend major revision rather than rejection. I would also ask the editor to ensure that the revised version includes the per-pixel uncertainty products, since the public release of the maps without them makes independent use difficult."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a genuine first — combining single-antenna and aperture-synthesis Stokes Q/U per frequency channel to make diffuse RM maps with sensitivity from degree to arcminute scales. The paper is honest about its limits, the method is documented in enough detail to reproduce, and the data products are released. It deserves peer review and probably acceptance with minor revisions.\n\nWhat's new: L10 did the combine for total and polarized intensity at one effective channel; previous diffuse RMs came from ST-only or single-antenna-only data. No one had done the channel-by-channel combine and then per-pixel RM fits. The power spectra and the comparisons to ST-only, single-antenna-only, compact source RMs, and the full-band GMIMS-HBN Faraday depths give the reader a real sense of what the combined map adds. The Sh2-216, IC 443, and bow-tie examples are useful case studies showing where the single-antenna component changes the interpretation.\n\nSoft spots, in rough order. First, the 35 MHz bandwidth. The authors say it themselves: they cannot detect Faraday complexity, and Figure 21c shows the four-channel RMs are systematically larger than the full-band Faraday depths. That means the absolute RM values in these maps should not be used as a catalog; the paper mostly uses them to trace structure, which is the right use. But the abstract's 'for magnetic field studies' could mislead a casual reader into treating these as final RMs. That is a presentation issue more than a technical one.\n\nSecond, the uv-gap validation is not as strong as it looks. Section 3.4.2 simulates the single-antenna data by convolving the true image with the same 9 m Gaussian beam model that is later deconvolved, so the test cannot catch an error in the beam model itself. It also excludes bright-source artifacts, which the authors acknowledge. A systematic error in the degree-scale Stokes Q/U would propagate coherently into all four channels and bias the RMs, not just add noise. The paper does show the ratio distributions with a wide spread and notes noise amplification beyond ~15 m, which is honest. But there is no external dataset that actually bridges the gap (L10 had Effelsberg 100 m; this study does not), so the claim that the missing spacings are harmless rests on the simulation alone. This is a real soft spot, but it is a soft spot in a demonstration, not in a claimed catalog.\n\nThird, no per-pixel RM error maps; the quoted RM uncertainties come from the linear fit scatter only, not from the systematic terms. Minor, given the paper's stated scope.\n\nWho this is for: anyone planning GMIMS-plus-interferometer combinations, and Galactic magnetism people who want to see what full-scale RM structure looks like before the broadband surveys arrive. A solid subfield contribution, not a paradigm shift. Recommendation: send it to a competent referee. The experiment is well-documented, the data are released, and the limitations are stated. The referee should push on the beam-model validation and the abstract wording, but the paper should not be desk-rejected.","headline":"Genuinely new full-scale diffuse RM maps, honestly limited; send to a serious referee, and push on the beam-model validation and the abstract's wording.","tokens_in":27588,"tokens_out":4723,"would_cite":true,"duration_ms":40881,"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":"The paper produces the first full-scale diffuse Galactic Faraday rotation maps, combining single-antenna GMIMS-HBN and aperture-synthesis CGPS polarization data to reach 3-arcmin resolution across all spatial scales.","keywords":["Faraday rotation","rotation measure","diffuse Galactic synchrotron emission","polarization","single-antenna plus interferometry","uv-plane feathering","Galactic magnetic field","GMIMS-HBN and CGPS"],"falsifier":"Compare the full-scale per-pixel RMs against the peak Faraday depth of the full-band GMIMS-HBN RM-synthesis cube at the same lines of sight: a systematic offset or sign disagreement beyond the stated uncertainties would show that the 35 MHz linear-fit RMs are not tracing the same magnetoionic structure. A second check is to recompute the RMs after perturbing the feathering boundaries, for example from 7 m to 20 m, and see whether the large-scale RM structures shift by more than the quoted errors.","tokens_in":26455,"feed_emoji":"📡","tokens_out":8205,"duration_ms":79222,"temperature":0.7,"pith_summary":"This paper reports the first experiment to combine single-antenna and aperture-synthesis polarization data over multiple frequency channels, producing diffuse Galactic synchrotron Faraday rotation maps that cover all spatial scales down to 3 arcmin. The authors merge GMIMS-HBN single-dish Stokes Q and U maps with CGPS aperture-synthesis data in four frequency bands, feathering the two datasets in the uv-plane and fitting per-pixel rotation measures to polarization angle versus $\\lambda^2$. The point of the experiment is to show that smooth, large-scale polarized emission, which interferometers filter out, can be recovered by adding single-antenna data, while the interferometer supplies arcminute detail. The resulting RM maps reveal both large-scale magnetic-field structures and small-scale RM variability, and demonstrate that useful diffuse-emission rotation measures can be extracted even from the narrow 35 MHz CGPS bandwidth, with the caveat that the RM values are sensitive to Faraday complexity.","feed_headline":"First all-scale Faraday rotation maps reach 3-arcmin resolution","feed_subtitle":"GMIMS-HBN single-dish plus CGPS interferometry maps diffuse Galactic magnetic fields from degree to arcminute scales.","key_machinery":"The central mechanism is feathering in the uv-plane: after Fourier transforming both datasets, the GMIMS-HBN visibilities are deconvolved by the single-antenna beam transform and low-pass filtered, then combined with the CGPS aperture-synthesis visibilities through complementary cubic weighting functions over the baseline range 8.572-17.144 m. This produces combined Stokes Q and U maps that keep sensitivity from the largest spatial scales down to the 3 arcmin resolution of the convolved maps; the per-pixel rotation measure is then obtained from a linear fit to polarization angle versus $\\lambda^2$ across the four 7.5 MHz channels.","core_discovery":"The central claim is that diffuse Galactic synchrotron emission can be mapped in Faraday rotation across all spatial scales down to 3 arcmin resolution for the first time, by combining GMIMS-HBN single-antenna and CGPS aperture-synthesis Stokes Q and U data after spatial filtering. The combination is carried out by deconvolving the single-antenna beam (modeled as a Gaussian with 9 m HWHM in the uv-plane), low-pass filtering the single-antenna visibilities at 18 m, and feathering with the aperture-synthesis visibilities using complementary cubic weights over baselines from 8.572 m to 17.144 m. Each of the four CGPS frequency channels receives a matching GMIMS-HBN band, and a linear fit of polarization angle versus $\\lambda^2$ gives an RM per pixel. A mock-observation simulation indicates that the missing intermediate uv spacings introduce only about 6% error in Stokes U away from bright sources, which the authors argue is adequate for a prototype demonstration. They show regions such as Sh2-216, IC 443, and the $\\ell = 173^\\circ$ H II complex where adding the single-antenna component changes the RM sign or reveals coherent structures invisible to aperture synthesis alone.","pith_inferences":["The paper's longitude-binned comparison hints that the 26 m single-antenna component dominates the combined RM pattern; a clean test would be to recompute the full-scale RMs with the single-antenna short spacings explicitly masked, to see whether the aperture-synthesis-only longitude trend reappears.","The success of the four-channel linear fit opens the door to recombining historical narrow-band polarization datasets wherever single-antenna and interferometric observations overlap in frequency, potentially extending full-scale RM coverage backward in time.","A natural extension is to apply the same method at lower frequencies, where Faraday depth sensitivity is higher but Faraday complexity is more severe, testing whether the narrow-band RM caveat becomes prohibitive.","If the simulated 6% error holds away from bright sources, the full-scale RM maps could also serve as priors for RM-synthesis cleaning of broadband data, helping to constrain Faraday complexity along each line of sight."],"forward_implications":["Galactic magnetic-field studies can now use diffuse-emission rotation measures at arcminute resolution rather than relying only on sparse compact-source RMs or single-antenna-only beams.","Smooth polarized regions that were invisible to aperture-synthesis-only surveys become measurable, while small-scale RM structure is preserved.","The same feathering recipe can be applied to future broadband combinations of single-antenna and interferometric polarization surveys, where wider frequency coverage should also resolve the Faraday complexity that the 35 MHz bandwidth cannot.","RM values derived from the narrow bandwidth must be treated as narrow-band estimates; the paper's comparison with the full GMIMS-HBN Faraday depth cube shows systematic magnitude differences that future broadband surveys will quantify.","Regions around bright sources such as Cygnus X, Cassiopeia A, and W3 carry instrumental-polarization artifacts and should be excluded until better leakage corrections are developed."],"supporting_citations":[{"why":"Supplies the CGPS aperture-synthesis polarization data, the feathering method used to combine single-antenna and interferometer maps, and the earlier full-band combination this work extends to four channels.","marker":"Landecker et al. 2010 (L10)"},{"why":"Provides the GMIMS-HBN single-antenna Stokes Q and U cube and the published Faraday depth cube used for comparison with the full-scale RMs.","marker":"Wolleben et al. 2021"},{"why":"Establishes that diffuse-emission RMs from CGPS aperture-synthesis data alone can trace large-scale magnetic-field structure, motivating the addition of single-antenna data.","marker":"Ordog et al. 2017"},{"why":"Shows the aperture-synthesis-only RM patterns and documents how narrow-band linear-fit RMs can deviate from a Burn slab, setting the cautionary framework for interpreting the combined maps.","marker":"Ordog et al. 2019"},{"why":"Provides the Burn slab model under which a linear fit of polarization angle versus $\\lambda^2$ over a narrow band still yields a meaningful mean Faraday depth.","marker":"Burn 1966"},{"why":"Defines RM synthesis, the broadband method the four-channel dataset cannot fully exploit, and provides the comparison standard for Faraday depth spectra.","marker":"Brentjens & de Bruyn 2005"},{"why":"Documents the incorrect polarization and RM results that can arise when large spatial scales are missing, the core motivation for combining the two datasets.","marker":"Gaensler et al. 2001"}],"fun_headline_variants":["First all-scale Faraday maps resolve to 3 arcmin","GMIMS+CGPS deliver all-scale RM maps at 3 arcmin","All-scale Galactic Faraday rotation maps from combined telescopes","Diffuse synchrotron Faraday maps span all scales to 3 arcmin","Single-dish plus interferometry maps RM at every scale"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the beam-deconvolved GMIMS-HBN visibilities faithfully represent the true large-scale polarized sky and that the feathering weights bridge the missing 8.6-17 m baselines without bias, since the only simulation validating this stitch excludes the bright-source regions that contaminate the real data.","fun_headline_variants_meta":{"raw":{"variants":["First all-scale Faraday maps resolve to 3 arcmin","GMIMS+CGPS deliver all-scale RM maps at 3 arcmin","All-scale Galactic Faraday rotation maps from combined telescopes","Diffuse synchrotron Faraday maps span all scales to 3 arcmin","Single-dish plus interferometry maps RM at every scale"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000344,"raw_usage":{"total_tokens":1971,"prompt_tokens":1111,"completion_tokens":860,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":727,"completion_tokens_details":{"reasoning_tokens":782}},"tokens_in":727,"tokens_out":860,"duration_ms":8712,"temperature":1.0,"reasoning_tokens":782,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:01:48.703098+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the full-scale per-pixel RMs against the peak Faraday depth of the full-band GMIMS-HBN RM-synthesis cube at the same lines of sight: a systematic offset or sign disagreement beyond the stated uncertainties would show that the 35 MHz linear-fit RMs are not tracing the same magnetoionic structure. A second check is to recompute the RMs after perturbing the feathering boundaries, for example from 7 m to 20 m, and see whether the large-scale RM structures shift by more than the quoted errors.","supporting_citations":[{"cited_title":"L., Reich, W., Reid, R","cited_arxiv_id":null,"evidence_quote":"Supplies the CGPS aperture-synthesis polarization data, the feathering method used to combine single-antenna and interferometer maps, and the earlier full-band combination this work extends to four channels."}],"review_version":1}