{"id":"d8aba408-7e73-4cbd-b9d3-04009465e3d9","arxiv_id":"2506.04818","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A 16-event single-active-region study finds stronger dimming-flare correlations than in the general population, and shows that LASCO-only CME speeds weaken dimming-speed correlations.","lead":"This paper tracks the temporary dark regions, called coronal dimmings, that a hyperactive sunspot region carved into the Sun's corona during the May 2024 storms, and links their size, brightness, and magnetic flux to the flares and CMEs that caused them. It reports that these dimmings correlate more tightly with flare size and CME speed than the general solar population, and argues that coronagraphs miss early CME acceleration that dimmings reveal.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Claimed enhancement over KD18 is not controlled for the Eq. 2 selection threshold, which is applied only to the May 2024 sample; the comparative correlations may reflect differential selection.","rationale":"The paper is a careful observational study with identical detection software, transparent parameter definitions, and a valuable single-AR dataset. The reader's conditional verdict is appropriate. My stress-test focuses on the single comparison that carries the abstract claim. The selection threshold is not merely a nuisance: it is chosen from the same data after visual classification, and it is not applied to the comparison sample. This makes the reported enhancement vulnerable to a differential selection bias whose direction is not controlled. The concern is not that the authors did anything improper; they state the preselection step plainly. The issue is that the central quantitative claim — 'stronger correlation than anticipated' — has not been shown to survive an apples-to-apples selection. A concrete re-analysis of KD18 with the same threshold, plus a formal difference-of-correlations test, would settle it. I therefore keep the reader's CONDITIONAL verdict and do not escalate to rejection; the dataset and methods remain useful regardless of the outcome of this test.","tokens_in":32370,"tokens_out":8371,"duration_ms":97080,"concrete_test":"Apply the same selection rule of Eq. (2) to the KD18 sample: recompute all flare-dimming correlations shown in Figs. 5–7 and Fig. E.1 using only KD18 events with mean first-hour area growth rate above 1.8e6 km2/s. Also run a permutation or Fisher z-test for the difference between the May 2024 and KD18 correlation coefficients. If, after matching the selection, the May 2024 correlations are no longer significantly larger than KD18, the claimed enhancement is a selection artifact rather than single-AR physics; if the differences persist, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is explicitly comparative: dimming parameters, especially area and area growth rate, correlate more strongly with GOES peak flux and fluence for AR 13664 than in the KD18 general population. This comparison is only valid if both samples are selected by equivalent criteria. They are not. In Sect. 3.3, the authors choose Eq. (2) — mean first-hour area growth rate > 1.8e6 km2/s — after inspecting superposed epoch curves of the same 67 flares (Fig. 4), and apply it to the May 2024 events only. The KD18 sample was selected by a different visual pre-selection procedure and includes B/C-class flares. Because the threshold truncates on area growth rate, one of the very parameters claimed to show enhanced correlation, the difference in correlation coefficients between the samples could be a selection effect. The direction of the bias is not obvious a priori: removing false detections can sharpen correlations, while range restriction on the predictor can attenuate them. Either way, the comparison is uncontrolled. In addition, the paper reports bootstrap error bars but never performs a formal test of the difference between correlation coefficients; for some headline pairs the intervals overlap, e.g., A vs F_T: 0.68 ± 0.10 for May 2024 versus 0.67 ± 0.06 for KD18. The authors acknowledge in Sect. 5.3 that 'a preselection step was necessary,' but that does not quantify the effect on the comparison. This is the load-bearing weakness for the main abstract claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a systematic study of coronal dimmings associated with flares and CMEs from active region 13664 during 2024 May 1–15. The authors identify 67 M- and X-class flares, 23 CMEs, and 22 dimming detections (16 on-disc), and derive characteristic dimming parameters (area, area growth rate, magnetic flux, brightness drop) using the Dissauer et al. detection and parameterization framework. They compare correlations between dimming parameters and GOES flare peak flux, fluence, and CME maximum velocity with the statistical sample of Dissauer et al. (2018b, 2019), finding enhanced correlations for the single-AR sample and arguing that LASCO-based CME velocities underestimate the true correlations because they miss the early acceleration phase. The central claims are that dimming area and area growth rate are tighter proxies for flare energy release in a single active region than in the general population, and that coronal dimmings capture low-corona CME dynamics better than coronagraphs.","tokens_in":32634,"tokens_out":6654,"duration_ms":66965,"significance":"If the central claims hold, the paper provides quantitative support for using coronal dimming area and growth rate as proxies for flare energy release and as early indicators of CME speed, with potential application to stellar CME studies. The single-AR sample is a valuable complement to the multi-AR statistical studies, and the careful uncertainty estimation via threshold variation and bootstrapping is a strength. However, the significance is moderated by two load-bearing concerns: the dimming sample is selected by a threshold derived from the same dataset, and the comparison with the KD18 sample is not controlled for that differential selection; additionally, the tables that are needed to reproduce the correlations contain inconsistent event numbering. These issues must be resolved before the claimed enhancements can be considered established.","major_comments":[{"comment":"The selection threshold on the first-hour mean area growth rate (Eq. 2) is derived from a superposed epoch analysis of the very same 67 flares (Fig. 4) and is then applied to that same sample to define the 16 on-disc dimmings. The KD18 comparison sample was selected by a different visual pre-selection procedure and includes B/C-class flares. Since the threshold truncates the sample on the area growth rate, the reported enhancements of the correlations involving A and \\dot{A} relative to KD18 could be partly a selection artifact. The paper acknowledges in Sect. 5.3 that 'a preselection step was necessary', but it does not quantify the effect on the comparison. Please add a robustness test, for example by applying an analogous growth-rate threshold to the KD18 sample (if the time series allow) or by recomputing the KD18 correlations on the M1.0+ subsample, and by reporting the correlations obtained with and without the threshold for the May 2024 full detection set.","section":"Section 3.3, Eq. (2)"},{"comment":"The abstract claims stronger correlations with both GOES peak flux and fluence. For the headline pair A vs F_T, the evidence does not support an enhancement: the May 2024 correlation is c = 0.68 ± 0.10 versus c = 0.67 ± 0.06 for KD18, which are indistinguishable within the quoted uncertainties. The paper reports bootstrap error bars but never performs a formal test of the difference between the two correlation coefficients. Please add a formal comparison (e.g., Fisher z-transform or bootstrap of the difference) for all pairs where an enhancement is claimed, and state explicitly which differences are statistically significant and which are not.","section":"Section 4.2, Fig. 5(b)"},{"comment":"The event numbering in Table B.1 does not match Table A.1. For example, Table B.1 lists N=4 as a dimming event on May 5 at 14:33 UT, but Table A.1 N=4 corresponds to a flare on May 5 at 09:23 UT with no dimming, while the 14:33 UT dimming is N=5 in Table A.1. Similar mismatches occur for other rows (e.g., B.1 N=27 appears to correspond to A.1 N=25). This inconsistency makes it impossible for a reader to associate the tabulated dimming parameters with the correct flare and CME properties, and casts doubt on the reproducibility of all correlation results that use these parameters. Please correct the tables and verify that the analysis code and figures used the correct event pairings.","section":"Tables A.1 and B.1"}],"minor_comments":[{"comment":"The caption uses 'pink' for the category that the text in Sect. 3.3 calls 'magenta'; please harmonize the terminology.","section":"Fig. 4 caption"},{"comment":"Typo: 'coronagrapo- hic' should be 'coronagraphic' in the sentence about the coronagraphic field of view.","section":"Section 4.4"},{"comment":"Minor grammatical issue: 'The data was rebinned' should be 'The data were rebinned'.","section":"Section 2.2"},{"comment":"The column header 'Flare Start' would be clearer as 'Flare Start (UT)'.","section":"Table B.1"},{"comment":"The reference to Veronig et al. (2025) as 'under review' should be updated if the Living Reviews article has been accepted or published by the time of publication.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The table-numbering inconsistency is the most serious technical issue; it suggests that the manuscript may have undergone a renumbering of events without a corresponding update of Table B.1. The authors should be asked to supply corrected tables and, ideally, a machine-readable data file for the 16 on-disc dimmings with unambiguous event identifiers. The selection-circularity concern is also central: the abstract's comparative claims cannot be evaluated without a controlled comparison or a clear statement that the enhancement is only relative to an uncontrolled historical sample."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe thing to know: this paper has a genuinely new dataset — a 14-day continuous, spatially resolved census of coronal dimmings from one active region (AR 13664) — and a new quantitative selection criterion (Eq. 2: mean area growth rate > 1.8×10^6 km^2/s within the first hour). That is real value. But the headline claim, that dimming parameters correlate more strongly with flare properties than in the general population, is not yet controlled for the way the sample was selected. The threshold in Eq. 2 was chosen after visually classifying the same 67 flares (Fig. 4) and is applied only to the May 2024 events; the KD18 comparison sample used a different pre-selection. So the difference in correlations could be a selection artifact, or a consequence of the M/X-only flare range, rather than something special about a single AR. The paper discusses this possibility but doesn't quantify it. For some headline pairs the bootstrap intervals overlap (e.g., A vs fluence: 0.68±0.10 vs 0.67±0.06 for KD18), and no formal test of the difference is given.\n\nWhat is solid: the data reduction is careful and transparent (threshold variation, bootstrapping, handling of the May 8 data gap). The LASCO-vs-STEREO re-analysis cleanly confirms that coronagraphs miss the early CME acceleration phase and thereby weaken dimming–velocity correlations — a practically relevant point for nowcasting. The within-AR correlations are credible.\n\nThe soft spot is the comparative claim. The authors are honest about the preselection step (Sect. 5.3), but they don't apply the same threshold to the KD18 sample or restrict it to M/X flares, and they don't test whether the correlation coefficients actually differ. That is the load-bearing weakness for the abstract's \"stronger correlation... than anticipated.\" The paper would be just as valuable with a more measured comparison, or with the formal analysis.\n\nThis deserves a serious referee. It's a good observational case study with a reusable tool and dataset; the main claim needs tempering or support by a matching analysis. For solar and space weather readers, it's worth engaging. I'd recommend peer review with a request for a proper comparison and a more careful interpretation.\n\nBest","headline":"A genuinely new single-AR dimming dataset and a reproducible selection tool, but the comparative claim of enhanced correlations over the general population is not yet controlled for the data-dependent threshold.","tokens_in":33200,"tokens_out":3944,"would_cite":true,"duration_ms":45682,"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 within a single active region, dimming area and area growth rate track flare peak flux and fluence more tightly than in the general population, and that coronagraph-only CME speeds underestimate the dimming–CME link.","keywords":["coronal dimmings","active region AR 13664","May 2024 solar storms","solar flares","coronal mass ejections","GOES soft X-ray flux","dimming area growth rate","EUV imaging"],"falsifier":"Take the 62-event comparison sample, apply the same selection used here—only flares of GOES class ≥M1.0 and only events whose first-hour mean area growth rate exceeds $1.8\\times10^{6}$ km$^{2}$ s$^{-1}$—and recompute the correlations of total dimming area with GOES peak flux and fluence; if the coefficients do not rise toward the paper's values ($c \\approx 0.78$ and $0.68$), the claimed single-AR enhancement is a selection artifact rather than a physical effect. Alternatively, a single visually unambiguous dimming associated with a fast CME that fails the growth-rate threshold would falsify the threshold's role as the discriminator of real dimmings.","tokens_in":32141,"feed_emoji":"☀️","tokens_out":11883,"duration_ms":117061,"temperature":0.7,"pith_summary":"Coronal dimmings are temporary darkenings of the Sun's corona in extreme-ultraviolet images, produced when plasma is evacuated during a coronal mass ejection. This paper argues that, within one hyperactive active region (AR 13664 during May 2024), the size of a dimming and how fast it grows are quantitative signatures of the energy released by the accompanying flare and of the speed the ejected plasma will reach. Across 16 on-disc dimmings, total dimming area and area growth rate correlate with GOES soft X-ray peak flux and fluence more strongly than in the general dimming population, because the events share the same magnetic environment. The paper also claims that coronagraph-only CME speeds, which miss the lower corona, systematically weaken the dimming–CME correlation; including EUV observations of the low corona restores it. If these claims hold, dimmings become a practical tool for estimating flare energy and early CME strength from EUV images alone, including for Earth-directed events.","feed_headline":"Coronal dimmings predict flare energy and CME speed","feed_subtitle":"In one active region, dimming area and growth rate track flare flux and CME speed; coronagraph-only speeds weaken it.","key_machinery":"The carrying object is the cumulative dimming mask built by logarithmic base-ratio thresholding of SDO/AIA 211 Å images at $\\log_{10}(I/I_0) \\le -0.19$, from which the paper derives the time evolution of dimming area $A(t)$, area growth rate $\\dot{A}(t)$, magnetic area $A_\\phi(t)$, unsigned magnetic flux $\\phi(t)$, and brightness drop $I_{\\rm drop}(t)$. The load-bearing selection rule is Eq. (2): a detection counts as a real coronal dimming only if its mean area growth rate during the first hour after flare onset exceeds $1.8\\times10^{6}$ km$^{2}$ s$^{-1}$, equivalent to $\\Delta A \\ge 6.48\\times10^{9}$ km$^{2}$; this threshold comes from a superposed-epoch analysis of events visually classified into clear, complicated, unclear, and absent dimmings. The threshold turns the 67 M/X-class flares into the 16 on-disc events whose log-log correlations with flare and CME parameters carry the paper's conclusions.","core_discovery":"Within a single active region the relation between dimmings and their parent flares is much tighter than in the wider dimming population, and the extra amount of correlation is physically informative rather than incidental. For the 16 on-disc dimmings from AR 13664, total dimming area $A$ correlates with GOES peak soft X-ray flux $F_P$ at $c = 0.78 \\pm 0.12$ (compared with $c = 0.53 \\pm 0.07$ for the comparison sample), and the magnetic dimming area $A_\\phi$ reaches $c = 0.84 \\pm 0.08$ with $F_P$; similar enhancement appears for flare fluence and for the area growth rate. The paper further shows that when CME maximum velocities are taken only from SOHO/LASCO coronagraphs the correlations with dimming parameters drop (for the comparison sample, from $c \\approx 0.56$–$0.69$ with EUV-inclusive velocities to $c \\approx 0.36$–$0.41$ with LASCO-only velocities), demonstrating that coronagraphs underestimate the dimming–CME link because they miss the early acceleration phase below about $2\\,R_\\odot$. It also finds that AR 13664's very strong magnetic fields suppress eruptions: only 23% of its M-class flares (and 83% of its X-class flares) were accompanied by CMEs, well below the general M-class association rate of roughly 60%.","pith_inferences":["Inference beyond the paper: part of the claimed enhancement over the comparison sample could be a selection effect, because the first-hour growth-rate threshold removes the slow-growing dimmings that would add scatter to the area–flux relation; applying the identical threshold to the comparison sample would test this.","Inference beyond the paper: if the single-AR correlations are physical, then global statistical samples are diluted by mixing different magnetic environments, and future dimming–flare correlations should be stratified by active-region magnetic flux or include region identity as a covariate.","Inference beyond the paper: the growth-rate threshold suggests an operational early-warning test—flag a ≥M1 flare once an EUV region grows by $6.48\\times10^{9}$ km$^{2}$ within an hour—and this could be validated in real time against LASCO CME speeds.","Inference beyond the paper: a direct check of the coronagraph-underestimation claim for the May 2024 events would be to compare LASCO-based $v_{\\max}$ correlations with velocities reconstructed from STEREO-A EUVI plus COR data, even at the limited 12° separation available."],"forward_implications":["Dimming area and area growth rate, measured from full-disc EUV images, can be used as quantitative proxies for GOES soft X-ray peak flux and fluence, at least for events from a single active region.","Magnetic dimming area and its growth rate can serve as early indicators of maximum CME speed, provided the CME kinematics include the low corona rather than only coronagraph data.","Coronagraph-only CME catalogues systematically weaken dimming–CME correlations because they miss the acceleration phase below about $2\\,R_\\odot$; future studies should combine EUV and coronagraphic measurements.","Flare–CME association rates are not universal: in AR 13664 only 23% of M-class flares and 83% of X-class flares had CMEs, reflecting strong magnetic confinement, so global rates should not be applied to individual active regions.","The same dimming diagnostics are relevant to stellar CME searches, where strong-field active regions may bias the detectability of stellar eruptions."],"supporting_citations":[{"why":"Supplies the region-growing dimming detection algorithm and the definitions of the characteristic dimming parameters used throughout.","marker":"Dissauer et al. (2018a)"},{"why":"Provides the 62-event general dimming population against which the May 2024 correlations and parameter ranges are compared.","marker":"Dissauer et al. (2018b)"},{"why":"Provides the STEREO-based CME velocities and dimming–velocity correlations that the paper re-derives from LASCO to demonstrate coronagraphic underestimation.","marker":"Dissauer et al. (2019)"},{"why":"Defines the CDAW SOHO/LASCO CME catalogue from which CME identifications and maximum velocities are taken.","marker":"Yashiro et al. (2004)"},{"why":"Gives the AR 13664 event overview and the list of major CMEs used to cross-check the flare–CME associations.","marker":"Hayakawa et al. (2025)"},{"why":"Supplies the general flare–CME association rates (about 60% for M-class flares) against which AR 13664's low eruptivity is measured.","marker":"Yashiro et al. (2006)"},{"why":"Supplies the correlation-strength categories and the flare ribbon–flux relation used to interpret the enhanced correlations.","marker":"Kazachenko et al. (2017)"},{"why":"Supports the magnetic-confinement explanation: large active regions with strong magnetic flux produce mostly confined flares.","marker":"Li et al. (2021)"},{"why":"Provides the expected probability of dimmings without CMEs, used to interpret the three dimmings that lack CME counterparts.","marker":"Veronig et al. (2021)"}],"fun_headline_variants":["Single-AR dimmings sharpen flare and CME correlations","Coronagraphs underestimate dimming–CME link by missing early phase","AR 13664 dimmings correlate tighter with flares than typical population","Strong fields suppress CMEs: only 23% of M-flares in AR 13664"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the visual classification of detections into clear, complicated, unclear, and absent dimmings is an unbiased ground truth, so that the derived threshold—an average area growth rate above $1.8\\times10^{6}$ km$^{2}$ s$^{-1}$ within the first hour—separates true plasma-depletion dimmings from unrelated EUV darkenings; if that premise fails, the selection of the 16 on-disc events and the correlations built on them become partly a selection artifact.","fun_headline_variants_meta":{"raw":{"variants":["Single-AR dimmings sharpen flare and CME correlations","Coronagraphs underestimate dimming–CME link by missing early phase","AR 13664 dimmings correlate tighter with flares than typical population","Strong fields suppress CMEs: only 23% of M-flares in AR 13664"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000244,"raw_usage":{"total_tokens":1679,"prompt_tokens":1241,"completion_tokens":438,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":857,"completion_tokens_details":{"reasoning_tokens":354}},"tokens_in":857,"tokens_out":438,"duration_ms":4808,"temperature":1.0,"reasoning_tokens":354,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T10:32:47.300688+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the 62-event comparison sample, apply the same selection used here—only flares of GOES class ≥M1.0 and only events whose first-hour mean area growth rate exceeds $1.8\\times10^{6}$ km$^{2}$ s$^{-1}$—and recompute the correlations of total dimming area with GOES peak flux and fluence; if the coefficients do not rise toward the paper's values ($c \\approx 0.78$ and $0.68$), the claimed single-AR enhancement is a selection artifact rather than a physical effect. Alternatively, a single visually unambiguous dimming associated with a fast CME that fails the growth-rate threshold would falsify the threshold's role as the discriminator of real dimmings.","supporting_citations":[{"cited_title":"M., Temmer , M., & Podladchikova , T","cited_arxiv_id":null,"evidence_quote":"Provides the STEREO-based CME velocities and dimming–velocity correlations that the paper re-derives from LASCO to demonstrate coronagraphic underestimation."},{"cited_title":"2004, Journal of Geophysical Research (Space Physics), 109, A07105","cited_arxiv_id":null,"evidence_quote":"Defines the CDAW SOHO/LASCO CME catalogue from which CME identifications and maximum velocities are taken."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the general flare–CME association rates (about 60% for M-class flares) against which AR 13664's low eruptivity is measured."},{"cited_title":"D., Lynch , B","cited_arxiv_id":null,"evidence_quote":"Supplies the correlation-strength categories and the flare ribbon–flux relation used to interpret the enhanced correlations."},{"cited_title":"M., Odert , P., Leitzinger , M., et al","cited_arxiv_id":null,"evidence_quote":"Provides the expected probability of dimmings without CMEs, used to interpret the three dimmings that lack CME counterparts."}],"review_version":1}